fix(compute): 按官网补对话型号的 tool_use 标签 (#119)

逐条对照供应商官方文档，核对 36 个对话规格与 70 个 Provider / 套餐预置条目（capabilities 缺 tool_use / tools / function_calling）的工具调用支持情况： - 补 tool_use 25 条：规格 9、随规格继承的接入条目 5、接入覆盖 11 - 保持原样 81 条：官方写明不支持或页面互相矛盾 10、已下线或数日内下线 26、官方当前文档无该型号或无工具调用说明 34、只在未适配接口上支持 2、官网读取失败 9 - 只改 capabilities，不动 routing；qwen-max 等按华北 2（北京）地域记，国际站不适用 - 同步单模型来源记录、供应商证据目录与来源缺口清单；逐条决定与依据见 docs/source-rechecks/2026-10-05-tool-calling/README.md - presetDataVersion 128 → 129
2026-10-22 05:43:42 +08:00 · 2026-10-04 18:26:40 -07:00
parent a94a751078
commit e8122c1d77
73 changed files with 1014 additions and 302 deletions
--- a/docs/model-sources/providers/alibaba/SOURCES.md
+++ b/docs/model-sources/providers/alibaba/SOURCES.md
@@ -35,10 +35,47 @@
 - 入口：[glm53-alibaba](https://www.alibabacloud.com/help/en/model-studio/glm-5-3)
 - 类型：`official-doc`；当前读取状态：`fetched`；地域/接入面：`阿里云国际托管 GLM-5.3；不作为智谱直连价证据`。

+### bailian-function-calling
+
+- 入口：[bailian-function-calling](https://help.aliyun.com/zh/model-studio/qwen-function-calling)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：百炼 Function Calling 指南的支持模型清单；清单按系列列出、不分地域，地域差异以型号页为准。
+
+### bailian-qwen-max-caps
+
+- 入口：[bailian-qwen-max-caps](https://help.aliyun.com/zh/model-studio/qwen-max)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：百炼 qwen-max 型号页模型能力表；华北 2（北京）支持 Function Calling，新加坡（国际）不支持。
+
+### bailian-qwen-plus-caps
+
+- 入口：[bailian-qwen-plus-caps](https://help.aliyun.com/zh/model-studio/qwen-plus)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：百炼 qwen-plus 型号页模型能力表；华北 2（北京）支持 Function Calling，新加坡、法兰克福、弗吉尼亚、中国香港不支持。
+
+### bailian-qwen3-vl-plus-caps
+
+- 入口：[bailian-qwen3-vl-plus-caps](https://help.aliyun.com/zh/model-studio/qwen3-vl-plus)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：百炼 qwen3-vl-plus 型号页模型能力表；华北 2（北京）支持 Function Calling，新加坡、法兰克福、弗吉尼亚、中国香港不支持。
+
+### bailian-qwen3-vl-flash-caps
+
+- 入口：[bailian-qwen3-vl-flash-caps](https://help.aliyun.com/zh/model-studio/qwen3-vl-flash)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：百炼 qwen3-vl-flash 型号页模型能力表；华北 2（北京）支持 Function Calling，新加坡、法兰克福、弗吉尼亚不支持。
+
+### bailian-glm-4-7-caps
+
+- 入口：[bailian-glm-4-7-caps](https://help.aliyun.com/zh/model-studio/glm-4-7)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：百炼托管 glm-4.7 型号页模型能力表；不作为智谱直连证据。
+
+### bailian-coding-plan-tools
+
+- 入口：[bailian-coding-plan-tools](https://help.aliyun.com/zh/model-studio/coding-plan)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：百炼 Coding Plan 概述；2026-10-05 只核对支持模型名单与适用的编程工具。
+
 ## 已知边界与核验说明

 Qwen3.7 Max 官网给出 1000000 上下文、131072 最大输出和北京 12/36 元价格，且明确纯文本。已合并两条冲突规格，并同步 API/Token Plan 输出值。Qwen Omni 是多模态输入、文本输出；图片价格按张。套餐支持名单单独引用 Token Plan 文档，不能由原厂型号存在推断套餐开放。

+2026-10-05 工具调用核对：以上 `bailian-function-calling`、`bailian-qwen-max-caps`、`bailian-qwen-plus-caps`、`bailian-qwen3-vl-plus-caps`、`bailian-qwen3-vl-flash-caps`、`bailian-glm-4-7-caps`、`bailian-coding-plan-tools` 只用于核对函数 / 工具调用是否受支持，读取日期 2026-10-05；没有借此复核窗口、价格等其他字段，也没有做账号调用验收。能力标签里只有 `tool_use` 一项以这些来源为依据。
+
 状态 `partial` 仅表示列出的字段已核实；不能推断整条配置、价格、账号权限或真实模型调用都已验收。`pending` 没有已核字段；`historical` 只作历史兼容参考。路由 tier/priority、产品标签和默认选择属于本仓策略，不伪装成官网数据。

 <!-- source-metadata:start -->
@@ -117,6 +154,76 @@ Qwen3.7 Max 官网给出 1000000 上下文、131072 最大输出和北京 12/36
      "checkedAt": "2026-10-03",
      "contentSha256": "10220b6d17ba456d7370f05953dff5fd9ee701d09a59002e87623221f15d3792",
      "resolvedUrl": "https://www.alibabacloud.com/help/en/model-studio/glm-5-3"
+    },
+    {
+      "id": "bailian-function-calling",
+      "url": "https://help.aliyun.com/zh/model-studio/qwen-function-calling",
+      "kind": "official-doc",
+      "scope": "百炼 Function Calling 指南的支持模型清单；清单按系列列出、不分地域，地域差异以型号页为准",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "eaa582b988cb485c01f2f76ca307c5874a1115fc4313be1aa009f44bc3aa17fd",
+      "resolvedUrl": "https://help.aliyun.com/zh/model-studio/qwen-function-calling"
+    },
+    {
+      "id": "bailian-qwen-max-caps",
+      "url": "https://help.aliyun.com/zh/model-studio/qwen-max",
+      "kind": "official-doc",
+      "scope": "百炼 qwen-max 型号页模型能力表；华北 2（北京）支持 Function Calling，新加坡（国际）不支持",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "a447759fba7a97afd75b9465cf45f7a8d90dfb26176bdc26c53a685688a0b857",
+      "resolvedUrl": "https://help.aliyun.com/zh/model-studio/qwen-max"
+    },
+    {
+      "id": "bailian-qwen-plus-caps",
+      "url": "https://help.aliyun.com/zh/model-studio/qwen-plus",
+      "kind": "official-doc",
+      "scope": "百炼 qwen-plus 型号页模型能力表；华北 2（北京）支持 Function Calling，新加坡、法兰克福、弗吉尼亚、中国香港不支持",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "cc82b2d5846d87b1745a628cdcf6613cb80c9e54a51dd1694515c7d35d64db1b",
+      "resolvedUrl": "https://help.aliyun.com/zh/model-studio/qwen-plus"
+    },
+    {
+      "id": "bailian-qwen3-vl-plus-caps",
+      "url": "https://help.aliyun.com/zh/model-studio/qwen3-vl-plus",
+      "kind": "official-doc",
+      "scope": "百炼 qwen3-vl-plus 型号页模型能力表；华北 2（北京）支持 Function Calling，新加坡、法兰克福、弗吉尼亚、中国香港不支持",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "193ff38139e8f7cf99fae00b67c8b5876c1c0df68011ae8fa0030e24d777d830",
+      "resolvedUrl": "https://help.aliyun.com/zh/model-studio/qwen3-vl-plus"
+    },
+    {
+      "id": "bailian-qwen3-vl-flash-caps",
+      "url": "https://help.aliyun.com/zh/model-studio/qwen3-vl-flash",
+      "kind": "official-doc",
+      "scope": "百炼 qwen3-vl-flash 型号页模型能力表；华北 2（北京）支持 Function Calling，新加坡、法兰克福、弗吉尼亚不支持",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "657d5c0c935568ad9153f3bd32ecc498de389cc2bb7614f9aa4dcc03f3a9777d",
+      "resolvedUrl": "https://help.aliyun.com/zh/model-studio/qwen3-vl-flash"
+    },
+    {
+      "id": "bailian-glm-4-7-caps",
+      "url": "https://help.aliyun.com/zh/model-studio/glm-4-7",
+      "kind": "official-doc",
+      "scope": "百炼托管 glm-4.7 型号页模型能力表；不作为智谱直连证据",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "b34aa5913f4a963f8580193b441af15baf6157d6f4a564263b26641d0f6765a9",
+      "resolvedUrl": "https://help.aliyun.com/zh/model-studio/glm-4-7"
+    },
+    {
+      "id": "bailian-coding-plan-tools",
+      "url": "https://help.aliyun.com/zh/model-studio/coding-plan",
+      "kind": "official-doc",
+      "scope": "百炼 Coding Plan 概述；2026-10-05 只核对支持模型名单与适用的编程工具",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "21bb14508f06ea365a249f572b652e6424dcc2055d3e49c55b4af2448d3fcec8",
+      "resolvedUrl": "https://help.aliyun.com/zh/model-studio/coding-plan"
    }
  ]
 }
--- a/docs/model-sources/providers/alibaba/models/glm-4.7.md
+++ b/docs/model-sources/providers/alibaba/models/glm-4.7.md
@@ -31,7 +31,8 @@
  "capabilities": [
    "chat",
    "reasoning",
-    "code"
+    "code",
+    "tool_use"
  ]
 }
 ```
@@ -41,7 +42,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `glm-4.7` | 202752 / 131072 | 套餐：未声明 / 未声明 | `modelName`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing) | partial | `ee9d1b36d9a055b5c3429e95261eeb0ad63ec3046a5c606c14ec147e6e6bf3e4` |
+| `glm-4.7` | 202752 / 131072 | 套餐：未声明 / 未声明 | `modelName`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing), [bailian-glm-4-7-caps](../SOURCES.md#bailian-glm-4-7-caps), [bailian-coding-plan-tools](../SOURCES.md#bailian-coding-plan-tools), [bailian-function-calling](../SOURCES.md#bailian-function-calling) | partial | `6176023da60e0822986ef3dd927fad84d1d1e07eef6c9af3b98fb7a860616945` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。百炼 glm-4.7 型号页模型能力表 Function Calling 支持；Coding Plan 概述的支持模型含 glm-4.7。百炼 Function Calling 指南注明调用 GLM 系列需在请求中传 tool_stream=true，否则不返回 tool_calls；套餐端点是否同样要求未核。依据：[bailian-glm-4-7-caps](../SOURCES.md#bailian-glm-4-7-caps)、[bailian-coding-plan-tools](../SOURCES.md#bailian-coding-plan-tools)、[bailian-function-calling](../SOURCES.md#bailian-function-calling)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ### compute/coding-plans/dashscope-token-plan.json

--- a/docs/model-sources/providers/alibaba/models/qwen-max.md
+++ b/docs/model-sources/providers/alibaba/models/qwen-max.md
@@ -32,7 +32,8 @@
    "reasoning",
    "code",
    "multilingual",
-    "long_context"
+    "long_context",
+    "tool_use"
  ],
  "spec.defaultTemperature": 0.7
 }
@@ -43,7 +44,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `qwen-max` | 32768 / 8192 | 非计价主数据 | `id`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing) | partial | `63f3d8412534f5a0fc5f00a68dd4db043e16d1597f164803f0a77b1e538709dc` |
+| `qwen-max` | 32768 / 8192 | 非计价主数据 | `id`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing), [bailian-qwen-max-caps](../SOURCES.md#bailian-qwen-max-caps), [bailian-function-calling](../SOURCES.md#bailian-function-calling) | partial | `0b0943664ad4fbcb94f1bf3e1f1e7cadd6187cdc93bef46015747317acf14a22` |
+
+- 2026-10-05 工具调用核对：`spec.capabilities` 增加 `tool_use`。百炼型号页模型能力表：华北 2（北京）Function Calling 支持，新加坡（国际）不支持；Function Calling 指南的支持清单列有 Qwen-Max 系列。标签按北京地域记，国际站接入不适用。依据：[bailian-qwen-max-caps](../SOURCES.md#bailian-qwen-max-caps)、[bailian-function-calling](../SOURCES.md#bailian-function-calling)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/alibaba/models/qwen-plus.md
+++ b/docs/model-sources/providers/alibaba/models/qwen-plus.md
@@ -31,7 +31,8 @@
    "chat",
    "reasoning",
    "code",
-    "thinking_mode"
+    "thinking_mode",
+    "tool_use"
  ],
  "spec.defaultTemperature": 0.7
 }
@@ -42,7 +43,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `qwen-plus` | 1000000 / 32768 | 非计价主数据 | `id`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing) | partial | `87b14206bafb51b5b464c85cd120f6c0f4d8fd393c4cd095a6247431f22a6726` |
+| `qwen-plus` | 1000000 / 32768 | 非计价主数据 | `id`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing), [bailian-qwen-plus-caps](../SOURCES.md#bailian-qwen-plus-caps), [bailian-function-calling](../SOURCES.md#bailian-function-calling) | partial | `a5fd883a10c7cb9e174dc802c0aac9fb050380ef064e235738cd6332e88f451d` |
+
+- 2026-10-05 工具调用核对：`spec.capabilities` 增加 `tool_use`。百炼型号页模型能力表：华北 2（北京）Function Calling 支持，新加坡、法兰克福、弗吉尼亚、中国香港不支持；Function Calling 指南的支持清单列有 Qwen-Plus 系列。标签按北京地域记，其他地域接入不适用。依据：[bailian-qwen-plus-caps](../SOURCES.md#bailian-qwen-plus-caps)、[bailian-function-calling](../SOURCES.md#bailian-function-calling)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/alibaba/models/qwen3-vl-flash.md
+++ b/docs/model-sources/providers/alibaba/models/qwen3-vl-flash.md
@@ -32,7 +32,8 @@
    "chat",
    "vision",
    "image_understanding",
-    "fast"
+    "fast",
+    "tool_use"
  ],
  "defaultTemperature": 0.7,
  "defaultTopP": 0.8,
@@ -46,7 +47,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `qwen3-vl-flash` | 262144 / 32768 | CNY：0.8 / 3 | `modelName`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing) | partial | `735e6fd04d6f3d60bdc440cca46565afbbfcabd3f27079e7a6f12255a45eca1f` |
+| `qwen3-vl-flash` | 262144 / 32768 | CNY：0.8 / 3 | `modelName`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing), [bailian-qwen3-vl-flash-caps](../SOURCES.md#bailian-qwen3-vl-flash-caps), [bailian-function-calling](../SOURCES.md#bailian-function-calling) | partial | `3d8dc40c45dd00e285c06d389fbf159b74d533093ad5a6c7ca65ac738a9c399f` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。随共享规格继承。本接入是华北 2（北京）兼容端点，型号页北京地域 Function Calling 支持。依据：[bailian-qwen3-vl-flash-caps](../SOURCES.md#bailian-qwen3-vl-flash-caps)、[bailian-function-calling](../SOURCES.md#bailian-function-calling)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ### compute/model-specs/qwen.json

@@ -66,7 +69,8 @@
    "chat",
    "vision",
    "image_understanding",
-    "fast"
+    "fast",
+    "tool_use"
  ],
  "spec.defaultTemperature": 0.7
 }
@@ -77,7 +81,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `qwen3-vl-flash` | 262144 / 32768 | 非计价主数据 | `id`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing) | partial | `92f4f43392b9b1c78a7a8555e75e1307140c8d8333b81a0eb15718130fa7db0e` |
+| `qwen3-vl-flash` | 262144 / 32768 | 非计价主数据 | `id`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing), [bailian-qwen3-vl-flash-caps](../SOURCES.md#bailian-qwen3-vl-flash-caps), [bailian-function-calling](../SOURCES.md#bailian-function-calling) | partial | `84e0c2757b3bf8fab93e4b3f462fb86dca0dc76abf2ccbbb761af5a23eb463bc` |
+
+- 2026-10-05 工具调用核对：`spec.capabilities` 增加 `tool_use`。百炼型号页模型能力表：华北 2（北京）Function Calling 支持，新加坡、法兰克福、弗吉尼亚不支持；Function Calling 指南的支持清单列有 Qwen3-VL-Flash 系列。标签按北京地域记，其他地域接入不适用。依据：[bailian-qwen3-vl-flash-caps](../SOURCES.md#bailian-qwen3-vl-flash-caps)、[bailian-function-calling](../SOURCES.md#bailian-function-calling)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/alibaba/models/qwen3-vl-plus.md
+++ b/docs/model-sources/providers/alibaba/models/qwen3-vl-plus.md
@@ -33,7 +33,8 @@
    "vision",
    "image_understanding",
    "ocr",
-    "chart_analysis"
+    "chart_analysis",
+    "tool_use"
  ],
  "defaultTemperature": 0.7,
  "defaultTopP": 0.8,
@@ -47,7 +48,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `qwen3-vl-plus` | 262144 / 32768 | CNY：1.5 / 6 | `modelName`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing) | partial | `51cc65a021a99eba029e212ea519c0d2a2e116d63f12522809ae51adba7126c0` |
+| `qwen3-vl-plus` | 262144 / 32768 | CNY：1.5 / 6 | `modelName`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing), [bailian-qwen3-vl-plus-caps](../SOURCES.md#bailian-qwen3-vl-plus-caps), [bailian-function-calling](../SOURCES.md#bailian-function-calling) | partial | `1109e206fb6c0925d0f7aaf9ad49aaee4af64d8549b2c7abb1ff19891d19b885` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。随共享规格继承。本接入是华北 2（北京）兼容端点，型号页北京地域 Function Calling 支持。依据：[bailian-qwen3-vl-plus-caps](../SOURCES.md#bailian-qwen3-vl-plus-caps)、[bailian-function-calling](../SOURCES.md#bailian-function-calling)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ### compute/model-specs/qwen.json

@@ -68,7 +71,8 @@
    "vision",
    "image_understanding",
    "ocr",
-    "chart_analysis"
+    "chart_analysis",
+    "tool_use"
  ],
  "spec.defaultTemperature": 0.7
 }
@@ -79,7 +83,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `qwen3-vl-plus` | 262144 / 32768 | 非计价主数据 | `id`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing) | partial | `d9fe3780a0249b4bacb26e4f63420359265c3f5f14526960c50d21292f5e233b` |
+| `qwen3-vl-plus` | 262144 / 32768 | 非计价主数据 | `id`→[qwen-pricing](../SOURCES.md#qwen-pricing) | [qwen-pricing](../SOURCES.md#qwen-pricing), [bailian-qwen3-vl-plus-caps](../SOURCES.md#bailian-qwen3-vl-plus-caps), [bailian-function-calling](../SOURCES.md#bailian-function-calling) | partial | `c7672b30f4e396e8f3635e12553d9df4edff9390d792e99acc1de8d133600a35` |
+
+- 2026-10-05 工具调用核对：`spec.capabilities` 增加 `tool_use`。百炼型号页模型能力表：华北 2（北京）Function Calling 支持，新加坡、法兰克福、弗吉尼亚、中国香港不支持；Function Calling 指南的支持清单列有 Qwen3-VL-Plus 系列。标签按北京地域记，其他地域接入不适用。依据：[bailian-qwen3-vl-plus-caps](../SOURCES.md#bailian-qwen3-vl-plus-caps)、[bailian-function-calling](../SOURCES.md#bailian-function-calling)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/baidu/SOURCES.md
+++ b/docs/model-sources/providers/baidu/SOURCES.md
@@ -15,10 +15,22 @@
 - 入口：[baidu-plan](https://cloud.baidu.com/doc/qianfan/s/imlg0beiu)
 - 类型：`official-doc`；当前读取状态：`fetched`；地域/接入面：`baidu 官方入口；具体地域与接入面待该页面逐字段确认`。

+### qianfan-function-calling
+
+- 入口：[qianfan-function-calling](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：千帆 Function calling 文档的支持模型范围（页面更新时间 2026-07-09）；千帆 V2 接口。
+
+### qianfan-coding-plan-tools
+
+- 入口：[qianfan-coding-plan-tools](https://cloud.baidu.com/doc/qianfan/s/imlg0beiu)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：千帆 Coding Plan 文档（页面更新时间 2026-06-03）；2026-10-05 只核对可配置的 Model Name 与适用工具，套餐已于 2026-07-13 停止新购。
+
 ## 已知边界与核验说明

 已登记官方入口并尝试读取。表中仅标为已核字段的部分有此次核对记录；其余存量参数待逐字段复核。

+2026-10-05 工具调用核对：以上 `qianfan-function-calling`、`qianfan-coding-plan-tools` 只用于核对函数 / 工具调用是否受支持，读取日期 2026-10-05；没有借此复核窗口、价格等其他字段，也没有做账号调用验收。能力标签里只有 `tool_use` 一项以这些来源为依据。
+
 状态 `partial` 仅表示列出的字段已核实；不能推断整条配置、价格、账号权限或真实模型调用都已验收。`pending` 没有已核字段；`historical` 只作历史兼容参考。路由 tier/priority、产品标签和默认选择属于本仓策略，不伪装成官网数据。

 <!-- source-metadata:start -->
@@ -47,6 +59,26 @@
      "checkedAt": "2026-10-03",
      "contentSha256": "a366e76ee3aeeecc787ba6d314076c20c1b960e03f47d82aef3cdb8b685ad595",
      "resolvedUrl": "https://cloud.baidu.com/doc/qianfan/s/imlg0beiu"
+    },
+    {
+      "id": "qianfan-function-calling",
+      "url": "https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5",
+      "kind": "official-doc",
+      "scope": "千帆 Function calling 文档的支持模型范围（页面更新时间 2026-07-09）；千帆 V2 接口",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "db0c016665da14a9a8739d7749ca349bbb4136d92030e380614dace1c2b9940b",
+      "resolvedUrl": "https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5"
+    },
+    {
+      "id": "qianfan-coding-plan-tools",
+      "url": "https://cloud.baidu.com/doc/qianfan/s/imlg0beiu",
+      "kind": "official-doc",
+      "scope": "千帆 Coding Plan 文档（页面更新时间 2026-06-03）；2026-10-05 只核对可配置的 Model Name 与适用工具，套餐已于 2026-07-13 停止新购",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "a366e76ee3aeeecc787ba6d314076c20c1b960e03f47d82aef3cdb8b685ad595",
+      "resolvedUrl": "https://cloud.baidu.com/doc/qianfan/s/imlg0beiu"
    }
  ]
 }
--- a/docs/model-sources/providers/baidu/models/ernie-x1.1.md
+++ b/docs/model-sources/providers/baidu/models/ernie-x1.1.md
@@ -33,7 +33,8 @@
    "reasoning",
    "deep_thinking",
    "math",
-    "code"
+    "code",
+    "tool_use"
  ],
  "inputPrice": 1,
  "outputPrice": 4
@@ -45,7 +46,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `ernie-x1.1` | 65536 / 65536 | CNY：1 / 4 | 待核实 | [baidu-models](../SOURCES.md#baidu-models) | pending | `39b6c69eacc5049a131c2b42bf54bc430531aac8dad426d88df12339cdff3e9e` |
+| `ernie-x1.1` | 65536 / 65536 | CNY：1 / 4 | 待核实 | [baidu-models](../SOURCES.md#baidu-models), [qianfan-function-calling](../SOURCES.md#qianfan-function-calling) | pending | `0e62f1467ecbb6f13c453ee3fdcc63b1ff50c718667f021f23bb8d45f81f4727` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。随共享规格继承。千帆 Function calling 文档「支持模型范围」的 ERNIE 系列列有 ernie-x1.1。依据：[qianfan-function-calling](../SOURCES.md#qianfan-function-calling)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ### compute/model-specs/baidu.json

@@ -66,7 +69,8 @@
    "reasoning",
    "deep_thinking",
    "math",
-    "code"
+    "code",
+    "tool_use"
  ],
  "spec.supportsReasoning": true,
  "spec.defaultTemperature": null
@@ -78,7 +82,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `ernie-x1.1` | 65536 / 65536 | 非计价主数据 | 待核实 | [baidu-models](../SOURCES.md#baidu-models) | pending | `0bfc1a2e9f52d3e8f1e3c9e70d4abf9a5c74d07e7a42750d2d232157191723f0` |
+| `ernie-x1.1` | 65536 / 65536 | 非计价主数据 | 待核实 | [baidu-models](../SOURCES.md#baidu-models), [qianfan-function-calling](../SOURCES.md#qianfan-function-calling) | pending | `bcc423916c3c70990d5f616a2c2c8ddf3c305c1d8f7a37334f096bd549022b69` |
+
+- 2026-10-05 工具调用核对：`spec.capabilities` 增加 `tool_use`。千帆 Function calling 文档「支持模型范围」的 ERNIE 系列列有 ernie-x1.1。依据：[qianfan-function-calling](../SOURCES.md#qianfan-function-calling)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/baidu/models/glm-5.md
+++ b/docs/model-sources/providers/baidu/models/glm-5.md
@@ -29,7 +29,8 @@
    "chat",
    "reasoning",
    "code",
-    "agent"
+    "agent",
+    "tool_use"
  ]
 }
 ```
@@ -39,7 +40,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `glm-5` | 未声明 / 未声明 | 套餐：未声明 / 未声明 | `modelName`→[baidu-models](../SOURCES.md#baidu-models) | [baidu-models](../SOURCES.md#baidu-models) | partial | `30124c9a2c4fd2f585b2076355bd65b0013232323456564c44787c78fc328c19` |
+| `glm-5` | 未声明 / 未声明 | 套餐：未声明 / 未声明 | `modelName`→[baidu-models](../SOURCES.md#baidu-models) | [baidu-models](../SOURCES.md#baidu-models), [qianfan-function-calling](../SOURCES.md#qianfan-function-calling), [qianfan-coding-plan-tools](../SOURCES.md#qianfan-coding-plan-tools) | partial | `3bf186dad57efece189eebe77031fe466294c7629acf51aa8a60435a187d8bcf` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。千帆 Function calling 文档「支持模型范围」列有 glm-5；Coding Plan 文档可配置的 Model Name 含 glm-5。套餐当前可用性未重新核实。依据：[qianfan-function-calling](../SOURCES.md#qianfan-function-calling)、[qianfan-coding-plan-tools](../SOURCES.md#qianfan-coding-plan-tools)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/baidu/models/kimi-k2.5.md
+++ b/docs/model-sources/providers/baidu/models/kimi-k2.5.md
@@ -30,7 +30,8 @@
    "reasoning",
    "code",
    "agent",
-    "long_context"
+    "long_context",
+    "tool_use"
  ]
 }
 ```
@@ -40,7 +41,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `kimi-k2.5` | 未声明 / 未声明 | 套餐：未声明 / 未声明 | `modelName`→[baidu-plan](../SOURCES.md#baidu-plan) | [baidu-plan](../SOURCES.md#baidu-plan) | partial | `00e6c83e5b65b32615ef20958e927961d36bd8302b6661fb5bc20cd84b6cf35a` |
+| `kimi-k2.5` | 未声明 / 未声明 | 套餐：未声明 / 未声明 | `modelName`→[baidu-plan](../SOURCES.md#baidu-plan) | [baidu-plan](../SOURCES.md#baidu-plan), [qianfan-function-calling](../SOURCES.md#qianfan-function-calling), [qianfan-coding-plan-tools](../SOURCES.md#qianfan-coding-plan-tools) | partial | `f16fd6976ebe4df6a68f7a350c8c69dfe8e4ab89ef5f9125c64cf370c83c0776` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。千帆 Function calling 文档「支持模型范围」列有 kimi-k2.5；Coding Plan 文档可配置的 Model Name 含 kimi-k2.5。套餐当前可用性未重新核实。依据：[qianfan-function-calling](../SOURCES.md#qianfan-function-calling)、[qianfan-coding-plan-tools](../SOURCES.md#qianfan-coding-plan-tools)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/baidu/models/minimax-m2.5.md
+++ b/docs/model-sources/providers/baidu/models/minimax-m2.5.md
@@ -28,7 +28,8 @@
  "capabilities": [
    "chat",
    "reasoning",
-    "code"
+    "code",
+    "tool_use"
  ]
 }
 ```
@@ -38,7 +39,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `minimax-m2.5` | 未声明 / 未声明 | 套餐：未声明 / 未声明 | `modelName`→[baidu-plan](../SOURCES.md#baidu-plan) | [baidu-plan](../SOURCES.md#baidu-plan) | partial | `703afe7b05fe28d702ddcded549439837f45108b77f36580318be80a6761546d` |
+| `minimax-m2.5` | 未声明 / 未声明 | 套餐：未声明 / 未声明 | `modelName`→[baidu-plan](../SOURCES.md#baidu-plan) | [baidu-plan](../SOURCES.md#baidu-plan), [qianfan-function-calling](../SOURCES.md#qianfan-function-calling), [qianfan-coding-plan-tools](../SOURCES.md#qianfan-coding-plan-tools) | partial | `059315e8a6596cd914c4d3e8a34e9fa8937550abf281e03519ec626a2e5eabbf` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。千帆 Function calling 文档「支持模型范围」列有 minimax-m2.5；Coding Plan 文档可配置的 Model Name 含 minimax-m2.5（套餐表标注即将下线）。套餐当前可用性未重新核实。依据：[qianfan-function-calling](../SOURCES.md#qianfan-function-calling)、[qianfan-coding-plan-tools](../SOURCES.md#qianfan-coding-plan-tools)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/ollama/SOURCES.md
+++ b/docs/model-sources/providers/ollama/SOURCES.md
@@ -11,10 +11,17 @@
 - 入口：[ollama](https://ollama.com/blog)
 - 类型：`official-doc`；当前读取状态：`fetched`；地域/接入面：`ollama 官方入口；具体地域与接入面待该页面逐字段确认`。

+### ollama-llama31-70b
+
+- 入口：[ollama-llama31-70b](https://ollama.com/library/llama3.1:70b)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：Ollama 官方模型库 llama3.1:70b 页；2026-10-05 只核对 tools 能力标记。
+
 ## 已知边界与核验说明

 本地模型是否安装由本机发现接口决定。官网产品动态只证明运行器更新，不能证明用户已安装新模型。

+2026-10-05 工具调用核对：以上 `ollama-llama31-70b` 只用于核对函数 / 工具调用是否受支持，读取日期 2026-10-05；没有借此复核窗口、价格等其他字段，也没有做账号调用验收。能力标签里只有 `tool_use` 一项以这些来源为依据。
+
 状态 `partial` 仅表示列出的字段已核实；不能推断整条配置、价格、账号权限或真实模型调用都已验收。`pending` 没有已核字段；`historical` 只作历史兼容参考。路由 tier/priority、产品标签和默认选择属于本仓策略，不伪装成官网数据。

 <!-- source-metadata:start -->
@@ -33,6 +40,16 @@
      "checkedAt": "2026-10-03",
      "contentSha256": "13d9d182a6ebedb79ceb57181c1471ee809cfd38b43cdc3accc783df9e70023e",
      "resolvedUrl": "https://ollama.com/blog"
+    },
+    {
+      "id": "ollama-llama31-70b",
+      "url": "https://ollama.com/library/llama3.1:70b",
+      "kind": "official-doc",
+      "scope": "Ollama 官方模型库 llama3.1:70b 页；2026-10-05 只核对 tools 能力标记",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "1621196d2098f28ef6e16f9320276e30b3c5363b237796cc1dbd76168094a259",
+      "resolvedUrl": "https://ollama.com/library/llama3.1:70b"
    }
  ]
 }
--- a/docs/model-sources/providers/ollama/models/llama3.1--70b.md
+++ b/docs/model-sources/providers/ollama/models/llama3.1--70b.md
@@ -30,7 +30,8 @@
  "capabilities": [
    "chat",
    "code",
-    "reasoning"
+    "reasoning",
+    "tool_use"
  ],
  "defaultTemperature": 0.8,
  "defaultTopP": 0.9
@@ -42,7 +43,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `llama3.1:70b` | 131072 / 8192 | USD：未声明 / 未声明 | 待核实 | [ollama](../SOURCES.md#ollama) | pending | `b66be25716689873b6ebbcd6863348791aa98452e41311478849a0c1e659c918` |
+| `llama3.1:70b` | 131072 / 8192 | USD：未声明 / 未声明 | 待核实 | [ollama](../SOURCES.md#ollama), [ollama-llama31-70b](../SOURCES.md#ollama-llama31-70b) | pending | `875fdb8c434149be5fa974c0baa22f6e2e5399a43f79ea45b98412e9c944b61f` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。Ollama 官方模型库 llama3.1:70b 页带 tools 能力标记。本地实际能力取决于用户拉取的权重与 Ollama 版本。依据：[ollama-llama31-70b](../SOURCES.md#ollama-llama31-70b)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/openai/SOURCES.md
+++ b/docs/model-sources/providers/openai/SOURCES.md
@@ -49,10 +49,32 @@
 - 入口：[gpt-5-4-nano-api](https://developers.openai.com/api/docs/models/gpt-5.4-nano.md)
 - 类型：`official-doc`；读取状态：`fetched`；适用范围：OpenAI 直连 API；USD 每百万 token，不能证明 Codex 订阅窗口或可用性。

+### openai-gpt-5-nano-tools
+
+- 入口：[openai-gpt-5-nano-tools](https://developers.openai.com/api/docs/models/gpt-5-nano.md)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：OpenAI 直连 API 型号页；2026-10-05 只核对 Supported features 是否含 function_calling 与端点支持情况。
+
+### openai-gpt-4-1-nano-tools
+
+- 入口：[openai-gpt-4-1-nano-tools](https://developers.openai.com/api/docs/models/gpt-4.1-nano.md)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：OpenAI 直连 API 型号页；2026-10-05 只核对 Supported features 是否含 function_calling 与端点支持情况。
+
+### openai-gpt-5-4-nano-tools
+
+- 入口：[openai-gpt-5-4-nano-tools](https://developers.openai.com/api/docs/models/gpt-5.4-nano.md)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：OpenAI 直连 API 型号页；2026-10-05 只核对 Supported features 是否含 function_calling 与端点支持情况。
+
+### openai-gpt-oss-120b-tools
+
+- 入口：[openai-gpt-oss-120b-tools](https://developers.openai.com/api/docs/models/gpt-oss-120b.md)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：OpenAI 开放权重型号页；直连只有 Responses 端点，其他托管平台另核；2026-10-05 只核对 function_calling。
+
 ## 已知边界与核验说明

 GPT-6.1 Sol 的窗口、输出、推理档位与直连计价已再次核对。API 参数不能自动套用到订阅后端；订阅的档位限制保留接入面记录。其他旧型号尚未在本次逐字段复核。

+2026-10-05 工具调用核对：以上 `openai-gpt-5-nano-tools`、`openai-gpt-4-1-nano-tools`、`openai-gpt-5-4-nano-tools`、`openai-gpt-oss-120b-tools` 只用于核对函数 / 工具调用是否受支持，读取日期 2026-10-05；没有借此复核窗口、价格等其他字段，也没有做账号调用验收。能力标签里只有 `tool_use` 一项以这些来源为依据。
+
 状态 `partial` 仅表示列出的字段已核实；不能推断整条配置、价格、账号权限或真实模型调用都已验收。`pending` 没有已核字段；`historical` 只作历史兼容参考。路由 tier/priority、产品标签和默认选择属于本仓策略，不伪装成官网数据。

 `extra.responsesOnly` 是为工具调用采用的保守协议收紧。官网要求工具调用走 Responses；普通无工具的 Chat Completions 仍支持，不能把此配置理解为原厂完全不支持 Chat Completions。
@@ -212,6 +234,46 @@ GPT-6.1 Sol 的窗口、输出、推理档位与直连计价已再次核对。AP
      "checkedAt": "2026-10-03",
      "contentSha256": "ac78c7daedd9543728f7a2c6ed453afce3bcbdcf590535fb4b65a7d2e7c43481",
      "resolvedUrl": "https://developers.openai.com/api/docs/deprecations.md"
+    },
+    {
+      "id": "openai-gpt-5-nano-tools",
+      "url": "https://developers.openai.com/api/docs/models/gpt-5-nano.md",
+      "kind": "official-doc",
+      "scope": "OpenAI 直连 API 型号页；2026-10-05 只核对 Supported features 是否含 function_calling 与端点支持情况",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "81508b2800ec4c9da22cf057ad4d6c2ba9237d98ec6a3915b5bc1df0ce45d131",
+      "resolvedUrl": "https://developers.openai.com/api/docs/models/gpt-5-nano.md"
+    },
+    {
+      "id": "openai-gpt-4-1-nano-tools",
+      "url": "https://developers.openai.com/api/docs/models/gpt-4.1-nano.md",
+      "kind": "official-doc",
+      "scope": "OpenAI 直连 API 型号页；2026-10-05 只核对 Supported features 是否含 function_calling 与端点支持情况",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "b7bd756c48509d7eaa8187100fc682537c78eec6c3ff099ad6d3e28ace1a86eb",
+      "resolvedUrl": "https://developers.openai.com/api/docs/models/gpt-4.1-nano.md"
+    },
+    {
+      "id": "openai-gpt-5-4-nano-tools",
+      "url": "https://developers.openai.com/api/docs/models/gpt-5.4-nano.md",
+      "kind": "official-doc",
+      "scope": "OpenAI 直连 API 型号页；2026-10-05 只核对 Supported features 是否含 function_calling 与端点支持情况",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "3dd20e2c95f09f3efb19c6255387a4e07d4d7c213367cca0a015e162a8f5b4e0",
+      "resolvedUrl": "https://developers.openai.com/api/docs/models/gpt-5.4-nano.md"
+    },
+    {
+      "id": "openai-gpt-oss-120b-tools",
+      "url": "https://developers.openai.com/api/docs/models/gpt-oss-120b.md",
+      "kind": "official-doc",
+      "scope": "OpenAI 开放权重型号页；直连只有 Responses 端点，其他托管平台另核；2026-10-05 只核对 function_calling",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "2b57ea5a172afd293be49e3acb3d270c060a31ade355f04a4b7c4f45d37fae72",
+      "resolvedUrl": "https://developers.openai.com/api/docs/models/gpt-oss-120b.md"
    }
  ]
 }
--- a/docs/model-sources/providers/openai/models/gpt-4.1-nano.md
+++ b/docs/model-sources/providers/openai/models/gpt-4.1-nano.md
@@ -32,7 +32,8 @@
    "chat",
    "code",
    "ultra_long_context",
-    "fast"
+    "fast",
+    "tool_use"
  ],
  "defaultTemperature": 1,
  "defaultTopP": 1,
@@ -46,7 +47,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `gpt-4.1-nano` | 1047576 / 32768 | USD：0.1 / 0.4 | 待核实 | [openai-models](../SOURCES.md#openai-models) | pending | `41339189c0645800f9be182f997c45ba1928d751f2d827a3767056c66c9bdd8f` |
+| `gpt-4.1-nano` | 1047576 / 32768 | USD：0.1 / 0.4 | 待核实 | [openai-models](../SOURCES.md#openai-models), [openai-gpt-4-1-nano-tools](../SOURCES.md#openai-gpt-4-1-nano-tools) | pending | `7b7fb17a5a0283edf7c7b5dc4e84ebd1f374b6fb3bbb55c46b2aeb0bae816c24` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。随共享规格继承。官网型号页 Supported features 列有 function_calling，Chat Completions 端点为 Supported。依据：[openai-gpt-4-1-nano-tools](../SOURCES.md#openai-gpt-4-1-nano-tools)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ### compute/model-specs/openai.json

@@ -66,7 +69,8 @@
    "chat",
    "code",
    "ultra_long_context",
-    "fast"
+    "fast",
+    "tool_use"
  ],
  "spec.defaultTemperature": 1
 }
@@ -77,7 +81,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `gpt-4.1-nano` | 1047576 / 32768 | 非计价主数据 | 待核实 | [openai-models](../SOURCES.md#openai-models) | pending | `10391c540c483d083065285db7767352c94a881f8cb6288fd590f5334bb3eed1` |
+| `gpt-4.1-nano` | 1047576 / 32768 | 非计价主数据 | 待核实 | [openai-models](../SOURCES.md#openai-models), [openai-gpt-4-1-nano-tools](../SOURCES.md#openai-gpt-4-1-nano-tools) | pending | `5bbaae9a35a6df5215ef9f38b1d7739131669c18ecdf3199e8f0a39a786505ee` |
+
+- 2026-10-05 工具调用核对：`spec.capabilities` 增加 `tool_use`。官网型号页 Supported features 列有 function_calling，Chat Completions 与 Responses 端点均为 Supported。依据：[openai-gpt-4-1-nano-tools](../SOURCES.md#openai-gpt-4-1-nano-tools)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/openai/models/gpt-5-nano.md
+++ b/docs/model-sources/providers/openai/models/gpt-5-nano.md
@@ -31,7 +31,8 @@
  "capabilities": [
    "chat",
    "code",
-    "fast"
+    "fast",
+    "tool_use"
  ],
  "inputPrice": 0.05,
  "outputPrice": 0.4
@@ -43,7 +44,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `gpt-5-nano` | 400000 / 128000 | USD：0.05 / 0.4 | 待核实 | [openai-models](../SOURCES.md#openai-models) | pending | `90af0d7d4b1a5df561d95e574e1ebc1c4c974d42175a6a1a3046135edb5f80b8` |
+| `gpt-5-nano` | 400000 / 128000 | USD：0.05 / 0.4 | 待核实 | [openai-models](../SOURCES.md#openai-models), [openai-gpt-5-nano-tools](../SOURCES.md#openai-gpt-5-nano-tools) | pending | `439b0250c82f3a9d0caeca732c4b993353e2ef8e98898dd4ed120a3f8f0fb79c` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。随共享规格继承。官网型号页 Supported features 列有 function_calling，Chat Completions 端点为 Supported。依据：[openai-gpt-5-nano-tools](../SOURCES.md#openai-gpt-5-nano-tools)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ### compute/model-specs/openai.json

@@ -62,7 +65,8 @@
  "spec.capabilities": [
    "chat",
    "code",
-    "fast"
+    "fast",
+    "tool_use"
  ],
  "spec.defaultTemperature": 1
 }
@@ -73,7 +77,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `gpt-5-nano` | 400000 / 128000 | 非计价主数据 | 待核实 | [openai-models](../SOURCES.md#openai-models) | pending | `bf4ea84c497783cfbb7eaf471b5b23f8799e85a8b8fd660c6832fa12fd390173` |
+| `gpt-5-nano` | 400000 / 128000 | 非计价主数据 | 待核实 | [openai-models](../SOURCES.md#openai-models), [openai-gpt-5-nano-tools](../SOURCES.md#openai-gpt-5-nano-tools) | pending | `662dfdfc1ce5bd7fff2097242c6f5087d8e2c0706d4c65aa3bcb8000e347f276` |
+
+- 2026-10-05 工具调用核对：`spec.capabilities` 增加 `tool_use`。官网型号页 Supported features 列有 function_calling，Chat Completions 与 Responses 端点均为 Supported。依据：[openai-gpt-5-nano-tools](../SOURCES.md#openai-gpt-5-nano-tools)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/openai/models/gpt-5.4-nano.md
+++ b/docs/model-sources/providers/openai/models/gpt-5.4-nano.md
@@ -33,7 +33,8 @@
    "reasoning",
    "code",
    "fast",
-    "long_context"
+    "long_context",
+    "tool_use"
  ],
  "inputPrice": 0.2,
  "outputPrice": 1.25,
@@ -55,7 +56,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `gpt-5.4-nano` | 400000 / 128000 | USD：0.2 / 1.25 | `modelName`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api)；`contextWindow`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api)；`maxOutputTokens`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api)；`inputPrice`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api)；`outputPrice`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api)；`extra.reasoning.supportedEfforts`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api)；`extra.cachedInputPrice`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api) | [gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api) | partial | `c9177352926f2671a78517659b30654fe41a14a2ebfcf435ba7854ee030bac0e` |
+| `gpt-5.4-nano` | 400000 / 128000 | USD：0.2 / 1.25 | `modelName`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api)；`contextWindow`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api)；`maxOutputTokens`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api)；`inputPrice`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api)；`outputPrice`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api)；`extra.reasoning.supportedEfforts`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api)；`extra.cachedInputPrice`→[gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api) | [gpt-5-4-nano-api](../SOURCES.md#gpt-5-4-nano-api), [openai-gpt-5-4-nano-tools](../SOURCES.md#openai-gpt-5-4-nano-tools) | partial | `821b813acdd4c3ab2a7d8ba25980dd7123d9735e25ac1de4cd9a7f7026197465` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。官网型号页 Supported features 列有 function_calling，Chat Completions 与 Responses 端点均为 Supported。依据：[openai-gpt-5-4-nano-tools](../SOURCES.md#openai-gpt-5-4-nano-tools)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 本轮读取该模型官方明文规格页，确认原厂直连窗口、输出、标准价格与推理合同。Pro 仅限 Responses；GPT-5.4 Pro 官方默认 medium，修正旧 high。none 在产品层映射 off，schema 不接受将 none 写成 defaultEffort，故未伪造其他显式默认档。

--- a/docs/model-sources/providers/openai/models/gpt-oss-120b.md
+++ b/docs/model-sources/providers/openai/models/gpt-oss-120b.md
@@ -30,7 +30,8 @@
  "spec.capabilities": [
    "chat",
    "code",
-    "reasoning"
+    "reasoning",
+    "tool_use"
  ]
 }
 ```
@@ -40,7 +41,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `gpt-oss-120b` | 128000 / 16384 | 非计价主数据 | 待核实 | [openai-models](../SOURCES.md#openai-models) | pending | `72d4d1688e4cccad25cb26f3fbaa100da07eff743b5b48dd1c4194221a996513` |
+| `gpt-oss-120b` | 128000 / 16384 | 非计价主数据 | 待核实 | [openai-models](../SOURCES.md#openai-models), [openai-gpt-oss-120b-tools](../SOURCES.md#openai-gpt-oss-120b-tools) | pending | `7b3895ebd43e215f2a8c9b0bbc968d698f3053df2dc24c8c756648de94b74762` |
+
+- 2026-10-05 工具调用核对：`spec.capabilities` 增加 `tool_use`。官网型号页 Supported features 列有 function_calling；OpenAI 直连只有 Responses 端点为 Supported（Chat Completions 为 Not supported），其他托管平台另核。依据：[openai-gpt-oss-120b-tools](../SOURCES.md#openai-gpt-oss-120b-tools)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/openrouter/SOURCES.md
+++ b/docs/model-sources/providers/openrouter/SOURCES.md
@@ -11,10 +11,17 @@
 - 入口：[openrouter-api](https://openrouter.ai/api/v1/models)
 - 类型：`official-api`；当前读取状态：`fetched`；地域/接入面：`OpenRouter 公共 Models API 快照；USD/token 换算为 USD/百万 token`。

+### openrouter-models-tools
+
+- 入口：[openrouter-models-tools](https://openrouter.ai/api/v1/models)
+- 类型：`official-api`；读取状态：`fetched`；适用范围：OpenRouter 官方模型目录 API；2026-10-05 只核对 openrouter/auto 的 supported_parameters。
+
 ## 已知边界与核验说明

 只使用平台自己的 Models API 解释该接入面的 ID、窗口和 token 价格。API 的价格单位是 USD/token，配置乘以 1000000；-1 表示动态路由占位，不代表免费。当前目录缺少 stealth/ox-alpha、openai/gpt-oss-120b:free、qwen/qwen3-coder:free，已从该 Provider 移除并添加 tombstones。缺少于目录只证明此次未列出，不推断原厂模型退役。

+2026-10-05 工具调用核对：以上 `openrouter-models-tools` 只用于核对函数 / 工具调用是否受支持，读取日期 2026-10-05；没有借此复核窗口、价格等其他字段，也没有做账号调用验收。能力标签里只有 `tool_use` 一项以这些来源为依据。
+
 状态 `partial` 仅表示列出的字段已核实；不能推断整条配置、价格、账号权限或真实模型调用都已验收。`pending` 没有已核字段；`historical` 只作历史兼容参考。路由 tier/priority、产品标签和默认选择属于本仓策略，不伪装成官网数据。

 <!-- source-metadata:start -->
@@ -33,6 +40,16 @@
      "checkedAt": "2026-10-03",
      "contentSha256": "17f5cbde8c7bbac639e17cabd2f90dcb82499e9b6bbff38d234ca0c7129b96c2",
      "resolvedUrl": "https://openrouter.ai/api/v1/models"
+    },
+    {
+      "id": "openrouter-models-tools",
+      "url": "https://openrouter.ai/api/v1/models",
+      "kind": "official-api",
+      "scope": "OpenRouter 官方模型目录 API；2026-10-05 只核对 openrouter/auto 的 supported_parameters",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "e417af0907c15237d9f86fd1eb6344902b50129aa38539c14a74244b1d11a199",
+      "resolvedUrl": "https://openrouter.ai/api/v1/models"
    }
  ]
 }
--- a/docs/model-sources/providers/openrouter/models/openrouter--auto.md
+++ b/docs/model-sources/providers/openrouter/models/openrouter--auto.md
@@ -29,7 +29,8 @@
  ],
  "capabilities": [
    "chat",
-    "auto_routing"
+    "auto_routing",
+    "tool_use"
  ],
  "defaultTemperature": 1,
  "defaultTopP": 1,
@@ -42,7 +43,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `openrouter/auto` | 2000000 / 16384 | USD：未声明 / 未声明 | `contextWindow`→[openrouter-api](../SOURCES.md#openrouter-api), `modelName`→[openrouter-api](../SOURCES.md#openrouter-api) | [openrouter-api](../SOURCES.md#openrouter-api) | partial | `f00c22f23fb9f8df157559d612b9c6d07a4df6959580b5da6eddee27b9aadf25` |
+| `openrouter/auto` | 2000000 / 16384 | USD：未声明 / 未声明 | `contextWindow`→[openrouter-api](../SOURCES.md#openrouter-api), `modelName`→[openrouter-api](../SOURCES.md#openrouter-api) | [openrouter-api](../SOURCES.md#openrouter-api), [openrouter-models-tools](../SOURCES.md#openrouter-models-tools) | partial | `d368331999e51b59f268f3405e5a948cd815fb471035b725e5b5d83648a2bc5f` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。OpenRouter 官方模型目录 API 中 openrouter/auto 的 supported_parameters 含 tools、tool_choice。实际能力取决于被路由到的后端模型。依据：[openrouter-models-tools](../SOURCES.md#openrouter-models-tools)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/tencent/SOURCES.md
+++ b/docs/model-sources/providers/tencent/SOURCES.md
@@ -11,10 +11,22 @@
 - 入口：[tencent](https://www.tencent.com/tencent-releases-and-open-sources-tencent-hy4-preview/)
 - 类型：`official-doc`；当前读取状态：`fetched`；地域/接入面：`tencent 官方入口；具体地域与接入面待该页面逐字段确认`。

+### tokenhub-model-list
+
+- 入口：[tokenhub-model-list](https://cloud.tencent.com/document/product/1823/130051)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：腾讯云 TokenHub 模型列表；2026-10-05 只核对各型号「能力支持」是否含 Function Calling。
+
+### tencent-token-plan-doc
+
+- 入口：[tencent-token-plan-doc](https://cloud.tencent.com/document/product/1823/130060)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：腾讯云 Token Plan 个人版套餐文档；2026-10-05 只核对可用模型名单、Model ID 与适配工具。
+
 ## 已知边界与核验说明

 已登记官方入口并尝试读取。表中仅标为已核字段的部分有此次核对记录；其余存量参数待逐字段复核。

+2026-10-05 工具调用核对：以上 `tokenhub-model-list`、`tencent-token-plan-doc` 只用于核对函数 / 工具调用是否受支持，读取日期 2026-10-05；没有借此复核窗口、价格等其他字段，也没有做账号调用验收。能力标签里只有 `tool_use` 一项以这些来源为依据。
+
 状态 `partial` 仅表示列出的字段已核实；不能推断整条配置、价格、账号权限或真实模型调用都已验收。`pending` 没有已核字段；`historical` 只作历史兼容参考。路由 tier/priority、产品标签和默认选择属于本仓策略，不伪装成官网数据。

 <!-- source-metadata:start -->
@@ -33,6 +45,26 @@
      "checkedAt": "2026-10-03",
      "contentSha256": "08770afd8e2b89d4c93be931d0a32b78e2eac1e97821ea86aa125305db5f085a",
      "resolvedUrl": "https://www.tencent.com/tencent-releases-and-open-sources-tencent-hy4-preview/"
+    },
+    {
+      "id": "tokenhub-model-list",
+      "url": "https://cloud.tencent.com/document/product/1823/130051",
+      "kind": "official-doc",
+      "scope": "腾讯云 TokenHub 模型列表；2026-10-05 只核对各型号「能力支持」是否含 Function Calling",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "ed3083e94bcf573f9b26d952197a7a1f363dc1a62295c50896d1a21d12abb8ae",
+      "resolvedUrl": "https://cloud.tencent.com/document/product/1823/130051"
+    },
+    {
+      "id": "tencent-token-plan-doc",
+      "url": "https://cloud.tencent.com/document/product/1823/130060",
+      "kind": "official-doc",
+      "scope": "腾讯云 Token Plan 个人版套餐文档；2026-10-05 只核对可用模型名单、Model ID 与适配工具",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "f6e23cdd02bf3733791e6aafc60e9e263ecd4f1680f42f4a9e828b6ebc795054",
+      "resolvedUrl": "https://cloud.tencent.com/document/product/1823/130060"
    }
  ]
 }
--- a/docs/model-sources/providers/tencent/models/minimax-m2.7.md
+++ b/docs/model-sources/providers/tencent/models/minimax-m2.7.md
@@ -29,7 +29,8 @@
    "chat",
    "reasoning",
    "code",
-    "agent"
+    "agent",
+    "tool_use"
  ]
 }
 ```
@@ -39,7 +40,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `minimax-m2.7` | 未声明 / 未声明 | 套餐：未声明 / 未声明 | 待核实 | [tencent](../SOURCES.md#tencent) | pending | `c1552bf2af76f109465fdae3de1d3141a6cef234fef7536eed3da12c58ed83f4` |
+| `minimax-m2.7` | 未声明 / 未声明 | 套餐：未声明 / 未声明 | 待核实 | [tencent](../SOURCES.md#tencent), [tokenhub-model-list](../SOURCES.md#tokenhub-model-list), [tencent-token-plan-doc](../SOURCES.md#tencent-token-plan-doc) | pending | `928432248506ad0475048b5757a3705f1396e9a3465a39aafe890655568dfa3c` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。TokenHub 模型列表 MiniMax-M2.7（minimax-m2.7）的能力支持含 Function Calling；Token Plan 文档的可用模型含 minimax-m2.7。依据：[tokenhub-model-list](../SOURCES.md#tokenhub-model-list)、[tencent-token-plan-doc](../SOURCES.md#tencent-token-plan-doc)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/tencent/models/tc-code-latest.md
+++ b/docs/model-sources/providers/tencent/models/tc-code-latest.md
@@ -28,7 +28,8 @@
  "capabilities": [
    "chat",
    "reasoning",
-    "code"
+    "code",
+    "tool_use"
  ]
 }
 ```
@@ -38,7 +39,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `tc-code-latest` | 未声明 / 未声明 | 套餐：未声明 / 未声明 | 待核实 | [tencent](../SOURCES.md#tencent) | pending | `04d8a41f82a0efdc743c71357a33552c60c838dd598ed10d99077173cad3ea67` |
+| `tc-code-latest` | 未声明 / 未声明 | 套餐：未声明 / 未声明 | 待核实 | [tencent](../SOURCES.md#tencent), [tencent-token-plan-doc](../SOURCES.md#tencent-token-plan-doc), [tokenhub-model-list](../SOURCES.md#tokenhub-model-list) | pending | `2d1b66426d85dad1936fac50182749e2858aa17d2d7acbb371ebe16ff6e8ce4f` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。Token Plan 文档把 tc-code-latest 列为 Auto 智能路由的 Model ID，套餐适配 Claude Code、OpenCode 等编程工具；TokenHub 模型列表里套餐当前可用型号的能力支持均含 Function Calling。别名本身没有单独的能力表，依据是套餐文档。依据：[tencent-token-plan-doc](../SOURCES.md#tencent-token-plan-doc)、[tokenhub-model-list](../SOURCES.md#tokenhub-model-list)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/volcengine/SOURCES.md
+++ b/docs/model-sources/providers/volcengine/SOURCES.md
@@ -23,10 +23,27 @@
 - 入口：[volcengine-seed](https://docs.volcengine.com/docs/82379/2549861)
 - 类型：`official-doc`；当前读取状态：`unreadable`；地域/接入面：`volcengine 官方入口；具体地域与接入面待该页面逐字段确认`。

+### volcengine-coding-get-started
+
+- 入口：[volcengine-coding-get-started](https://docs.volcengine.com/docs/ark/coding-plan-personal-get-started?lang=zh)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：方舟 Coding Plan 个人版快速开始（浏览器渲染后读取，摘要为渲染正文；页面更新时间 2026-09-23）。
+
+### volcengine-coding-overview
+
+- 入口：[volcengine-coding-overview](https://docs.volcengine.com/docs/ark/coding-plan-personal-plan-overview?lang=zh)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：方舟 Coding Plan 个人版套餐概览（浏览器渲染后读取，摘要为渲染正文；页面更新时间 2026-09-29）。
+
+### volcengine-models-rendered
+
+- 入口：[volcengine-models-rendered](https://docs.volcengine.com/docs/ark/model-list?lang=zh)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：方舟模型列表（浏览器渲染后读取，摘要为渲染正文；页面更新时间 2026-09-28）；2026-10-05 只核对「工具调用能力」表。
+
 ## 已知边界与核验说明

 官方页面本次返回 JavaScript 页面壳，不能把 HTTP 200 当参数证明。Seed 2.1 Pro/Turbo 当前配置的输入/输出数字与官网最新值需通过控制台或正文继续核实；本次不采用开发者社区文章补数。

+2026-10-05 工具调用核对：以上 `volcengine-coding-get-started`、`volcengine-coding-overview`、`volcengine-models-rendered` 只用于核对函数 / 工具调用是否受支持，读取日期 2026-10-05；没有借此复核窗口、价格等其他字段，也没有做账号调用验收。能力标签里只有 `tool_use` 一项以这些来源为依据。
+
 状态 `partial` 仅表示列出的字段已核实；不能推断整条配置、价格、账号权限或真实模型调用都已验收。`pending` 没有已核字段；`historical` 只作历史兼容参考。路由 tier/priority、产品标签和默认选择属于本仓策略，不伪装成官网数据。

 <!-- source-metadata:start -->
@@ -75,6 +92,36 @@
      "checkedAt": "2026-10-03",
      "contentSha256": null,
      "resolvedUrl": null
+    },
+    {
+      "id": "volcengine-coding-get-started",
+      "url": "https://docs.volcengine.com/docs/ark/coding-plan-personal-get-started?lang=zh",
+      "kind": "official-doc",
+      "scope": "方舟 Coding Plan 个人版快速开始（浏览器渲染后读取，摘要为渲染正文；页面更新时间 2026-09-23）",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "dbd6bc539e6671be232d086ff0c0a9d57314dd534b7151369184b2b40a89f461",
+      "resolvedUrl": "https://docs.volcengine.com/docs/ark/coding-plan-personal-get-started?lang=zh"
+    },
+    {
+      "id": "volcengine-coding-overview",
+      "url": "https://docs.volcengine.com/docs/ark/coding-plan-personal-plan-overview?lang=zh",
+      "kind": "official-doc",
+      "scope": "方舟 Coding Plan 个人版套餐概览（浏览器渲染后读取，摘要为渲染正文；页面更新时间 2026-09-29）",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "5fd7c037c21f6c8072abd7cbbc31078a7c22c7b2abd35f14826241c636ece4d4",
+      "resolvedUrl": "https://docs.volcengine.com/docs/ark/coding-plan-personal-plan-overview?lang=zh"
+    },
+    {
+      "id": "volcengine-models-rendered",
+      "url": "https://docs.volcengine.com/docs/ark/model-list?lang=zh",
+      "kind": "official-doc",
+      "scope": "方舟模型列表（浏览器渲染后读取，摘要为渲染正文；页面更新时间 2026-09-28）；2026-10-05 只核对「工具调用能力」表",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "5092413aad51eb2271bef9dfff5e756bd0d28e492efb0569a5311300a43616ba",
+      "resolvedUrl": "https://docs.volcengine.com/docs/ark/model-list?lang=zh"
    }
  ]
 }
--- a/docs/model-sources/providers/volcengine/models/ark-code-latest.md
+++ b/docs/model-sources/providers/volcengine/models/ark-code-latest.md
@@ -28,7 +28,8 @@
  "capabilities": [
    "code",
    "reasoning",
-    "chat"
+    "chat",
+    "tool_use"
  ]
 }
 ```
@@ -38,7 +39,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `ark-code-latest` | 未声明 / 未声明 | 套餐：未声明 / 未声明 | 待核实 | [volcengine-models](../SOURCES.md#volcengine-models) | pending | `be9dd39891377694e9dc4cd374dd621c7e95f522e01c6bb85fda4fe4ab005d8a` |
+| `ark-code-latest` | 未声明 / 未声明 | 套餐：未声明 / 未声明 | 待核实 | [volcengine-models](../SOURCES.md#volcengine-models), [volcengine-coding-get-started](../SOURCES.md#volcengine-coding-get-started), [volcengine-coding-overview](../SOURCES.md#volcengine-coding-overview), [volcengine-models-rendered](../SOURCES.md#volcengine-models-rendered) | pending | `46403787cddd2882c0efad76fffc3f355950f86f9d3472529eadcb1859a26fd8` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。Coding Plan 快速开始把 ark-code-latest 作为 Claude Code 的 ANTHROPIC_MODEL 配置；套餐概览写明套餐额度仅在 AI 编程工具中生效；方舟模型列表「工具调用能力」表里套餐内的豆包、GLM、DeepSeek 型号函数调用均为支持。别名本身没有单独的能力表，依据是套餐文档。依据：[volcengine-coding-get-started](../SOURCES.md#volcengine-coding-get-started)、[volcengine-coding-overview](../SOURCES.md#volcengine-coding-overview)、[volcengine-models-rendered](../SOURCES.md#volcengine-models-rendered)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/zhipu/SOURCES.md
+++ b/docs/model-sources/providers/zhipu/SOURCES.md
@@ -33,10 +33,27 @@
 - 入口：[glm52-intl-api](https://docs.z.ai/api-reference/llm/chat-completion.md)
 - 类型：`official-doc`；读取状态：`fetched`；适用范围：Z.AI 国际原厂 Chat API 的 GLM5.2 effort 合同；套餐实际账号权限另核。

+### zhipu-glm-4-6-page
+
+- 入口：[zhipu-glm-4-6-page](https://docs.bigmodel.cn/cn/guide/models/text/glm-4.6.md)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：智谱开放平台 GLM-4.6 型号页；2026-10-05 只核对「能力支持」中的 Function Calling。
+
+### zhipu-glm-4-7-page
+
+- 入口：[zhipu-glm-4-7-page](https://docs.bigmodel.cn/cn/guide/models/text/glm-4.7.md)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：智谱开放平台 GLM-4.7 系列页（含 GLM-4.7-FlashX 页签）；2026-10-05 只核对系列「能力支持」中的 Function Calling。
+
+### zhipu-model-overview
+
+- 入口：[zhipu-model-overview](https://docs.bigmodel.cn/cn/guide/start/model-overview.md)
+- 类型：`official-doc`；读取状态：`fetched`；适用范围：智谱开放平台模型概览；GLM-4.7-FlashX 条目指向 GLM-4.7 系列页。
+
 ## 已知边界与核验说明

 国内官方价格页确认 5.3/Flash/FlashX 为 8/28、0.8/2.8、2/7 元/百万 tokens。官网给出的窗口简写 1M/128K 未直接作为跨接入面字节级精确值证明；本次不改已有换算。低/高/max、默认 max 和始终思考已核对。Coding Plan 提供 5.3 与 Flash，不提供 FlashX；国际套餐端点的账号实际可用性仍需验证。

+2026-10-05 工具调用核对：以上 `zhipu-glm-4-6-page`、`zhipu-glm-4-7-page`、`zhipu-model-overview` 只用于核对函数 / 工具调用是否受支持，读取日期 2026-10-05；没有借此复核窗口、价格等其他字段，也没有做账号调用验收。能力标签里只有 `tool_use` 一项以这些来源为依据。
+
 状态 `partial` 仅表示列出的字段已核实；不能推断整条配置、价格、账号权限或真实模型调用都已验收。`pending` 没有已核字段；`historical` 只作历史兼容参考。路由 tier/priority、产品标签和默认选择属于本仓策略，不伪装成官网数据。

 <!-- source-metadata:start -->
@@ -105,6 +122,36 @@
      "checkedAt": "2026-10-03",
      "contentSha256": "b05ada5d25382772edafd1b12dd2ead1da430a8817b5b748f2fb8dee2d7a6c10",
      "resolvedUrl": "https://docs.z.ai/api-reference/llm/chat-completion.md"
+    },
+    {
+      "id": "zhipu-glm-4-6-page",
+      "url": "https://docs.bigmodel.cn/cn/guide/models/text/glm-4.6.md",
+      "kind": "official-doc",
+      "scope": "智谱开放平台 GLM-4.6 型号页；2026-10-05 只核对「能力支持」中的 Function Calling",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "07cdc1b271155ccf16d43a1c21b6c690e32c826c902fa92bac9f4cab367f53e6",
+      "resolvedUrl": "https://docs.bigmodel.cn/cn/guide/models/text/glm-4.6.md"
+    },
+    {
+      "id": "zhipu-glm-4-7-page",
+      "url": "https://docs.bigmodel.cn/cn/guide/models/text/glm-4.7.md",
+      "kind": "official-doc",
+      "scope": "智谱开放平台 GLM-4.7 系列页（含 GLM-4.7-FlashX 页签）；2026-10-05 只核对系列「能力支持」中的 Function Calling",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "25ce7e68aa5afe76ee9d768f2556ff0a9577020244f8edc42a599eec181e41ba",
+      "resolvedUrl": "https://docs.bigmodel.cn/cn/guide/models/text/glm-4.7.md"
+    },
+    {
+      "id": "zhipu-model-overview",
+      "url": "https://docs.bigmodel.cn/cn/guide/start/model-overview.md",
+      "kind": "official-doc",
+      "scope": "智谱开放平台模型概览；GLM-4.7-FlashX 条目指向 GLM-4.7 系列页",
+      "retrieval": "fetched",
+      "checkedAt": "2026-10-05",
+      "contentSha256": "0b00f14c6071ef69fe48956d1199f534652309d63e859fae04c602dd734c6466",
+      "resolvedUrl": "https://docs.bigmodel.cn/cn/guide/start/model-overview.md"
    }
  ]
 }
--- a/docs/model-sources/providers/zhipu/models/glm-4.6.md
+++ b/docs/model-sources/providers/zhipu/models/glm-4.6.md
@@ -72,7 +72,8 @@
    "reasoning",
    "code",
    "multilingual",
-    "deep_thinking"
+    "deep_thinking",
+    "tool_use"
  ],
  "spec.supportsReasoning": true,
  "spec.defaultTemperature": 1
@@ -84,7 +85,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `glm-4.6` | 200000 / 128000 | 非计价主数据 | `id`→[glm-pricing](../SOURCES.md#glm-pricing) | [glm-pricing](../SOURCES.md#glm-pricing) | partial | `d000dbabac9ef3f752bc3afb854a3e84655dd496591354f05ffd9e93c5f46233` |
+| `glm-4.6` | 200000 / 128000 | 非计价主数据 | `id`→[glm-pricing](../SOURCES.md#glm-pricing) | [glm-pricing](../SOURCES.md#glm-pricing), [zhipu-glm-4-6-page](../SOURCES.md#zhipu-glm-4-6-page) | partial | `fdc0405d98bbabb6f2afd934364f84f18d9da89bd6b7a0b5dd006b05e3c53623` |
+
+- 2026-10-05 工具调用核对：`spec.capabilities` 增加 `tool_use`。官网 GLM-4.6 型号页「能力支持」列有 Function Calling。依据：[zhipu-glm-4-6-page](../SOURCES.md#zhipu-glm-4-6-page)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/model-sources/providers/zhipu/models/glm-4.7-flashx.md
+++ b/docs/model-sources/providers/zhipu/models/glm-4.7-flashx.md
@@ -32,7 +32,8 @@
    "chat",
    "fast",
    "long_context",
-    "multilingual"
+    "multilingual",
+    "tool_use"
  ],
  "defaultTemperature": 1,
  "defaultTopP": 0.95,
@@ -47,7 +48,9 @@

 | 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
 | --- | --- | --- | --- | --- | --- | --- |
-| `glm-4.7-flashx` | 200000 / 131072 | CNY：0.5 / 3 | `modelName`→[glm-pricing](../SOURCES.md#glm-pricing) | [glm-pricing](../SOURCES.md#glm-pricing) | partial | `60dcdbc2f8c517e20d6b926d0c38c84fd6b139ce3b5331ad0dc5eb8c801e9509` |
+| `glm-4.7-flashx` | 200000 / 131072 | CNY：0.5 / 3 | `modelName`→[glm-pricing](../SOURCES.md#glm-pricing) | [glm-pricing](../SOURCES.md#glm-pricing), [zhipu-glm-4-7-page](../SOURCES.md#zhipu-glm-4-7-page), [zhipu-model-overview](../SOURCES.md#zhipu-model-overview) | partial | `1f1692079bcaa050d7c6ce699b527af6a215cca735befcf11b19e736e26538c6` |
+
+- 2026-10-05 工具调用核对：`capabilities` 增加 `tool_use`。官网 GLM-4.7 系列页含 GLM-4.7-FlashX 页签，系列「能力支持」列有 Function Calling；模型概览的 GLM-4.7-FlashX 条目指向该页。精确 API ID 仍未单独证实，绑定方式不变。依据：[zhipu-glm-4-7-page](../SOURCES.md#zhipu-glm-4-7-page)、[zhipu-model-overview](../SOURCES.md#zhipu-model-overview)。本次只核对工具调用一项，其余标签仍是本仓产品标签，核验状态不变。

 ## 下次更新核查

--- a/docs/source-gaps/README.md
+++ b/docs/source-gaps/README.md
@@ -5,8 +5,8 @@
 供应商官网入口均已登记；“没有找到供应商官网”目前没有已登记案例。能明确单列的是官网入口读取失败/页面壳、未取得型号专属规格或身份揭示证据，以及缺少字段级证明。不能把这些情况混为“没有官网”。

 - 未登记官方入口：**0 条**。
- 未取得可读官网证据：**87 条**。
- 入口有读取记录，但本条模型无字段证明：**185 条**。
+- 未取得可读官网证据：**86 条**。
+- 入口有读取记录，但本条模型无字段证明：**186 条**。
 - 仅确认 ID，参数来源未登记：**216 条**。
 - 已有部分参数证明，仍需区分未证明字段：**137 条**。
 - 历史身份/参数不证明当前可用性：**1 条**。
@@ -38,7 +38,7 @@
 | [stability](providers/stability.md) | 0 | 0 | 3 | 0 | 0 | 0 |
 | [stealth](providers/stealth.md) | 0 | 0 | 0 | 0 | 0 | 1 |
 | [tencent](providers/tencent.md) | 0 | 0 | 23 | 1 | 1 | 0 |
-| [volcengine](providers/volcengine.md) | 0 | 66 | 0 | 0 | 0 | 0 |
+| [volcengine](providers/volcengine.md) | 0 | 65 | 1 | 0 | 0 | 0 |
 | [xai](providers/xai.md) | 0 | 0 | 5 | 3 | 1 | 0 |
 | [xiaomi](providers/xiaomi.md) | 0 | 0 | 8 | 7 | 10 | 0 |
 | [xunfei](providers/xunfei.md) | 0 | 0 | 2 | 0 | 2 | 0 |
--- a/docs/source-gaps/providers/alibaba.md
+++ b/docs/source-gaps/providers/alibaba.md
@@ -41,7 +41,7 @@
 | [`deepseek-v4-flash-0731`](../../model-sources/providers/alibaba/models/deepseek-v4-flash-0731.md) | [`compute/coding-plans/dashscope-token-plan.json`](../../../compute/coding-plans/dashscope-token-plan.json) | `contextWindow`、`extra.reasoning.defaultEffort`、`extra.reasoning.supportedEfforts`、`maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
 | [`deepseek-v4-pro-0813`](../../model-sources/providers/alibaba/models/deepseek-v4-pro-0813.md) | [`compute/coding-plans/dashscope-token-plan.json`](../../../compute/coding-plans/dashscope-token-plan.json) | `contextWindow`、`extra.reasoning.defaultEffort`、`extra.reasoning.supportedEfforts`、`maxOutputTokens` | [qwen-models](../../model-sources/providers/alibaba/SOURCES.md#qwen-models)：fetched |
 | [`deepseek-v4.1-flash`](../../model-sources/providers/alibaba/models/deepseek-v4.1-flash.md) | [`compute/coding-plans/dashscope-token-plan.json`](../../../compute/coding-plans/dashscope-token-plan.json) | `contextWindow`、`extra.reasoning.defaultEffort`、`extra.reasoning.supportedEfforts`、`maxOutputTokens` | [qwen-models](../../model-sources/providers/alibaba/SOURCES.md#qwen-models)：fetched |
-| [`glm-4.7`](../../model-sources/providers/alibaba/models/glm-4.7.md) | [`compute/coding-plans/dashscope-coding.json`](../../../compute/coding-plans/dashscope-coding.json) | `contextWindow`、`maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
+| [`glm-4.7`](../../model-sources/providers/alibaba/models/glm-4.7.md) | [`compute/coding-plans/dashscope-coding.json`](../../../compute/coding-plans/dashscope-coding.json) | `contextWindow`、`maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched；[bailian-glm-4-7-caps](../../model-sources/providers/alibaba/SOURCES.md#bailian-glm-4-7-caps)：fetched；[bailian-coding-plan-tools](../../model-sources/providers/alibaba/SOURCES.md#bailian-coding-plan-tools)：fetched；[bailian-function-calling](../../model-sources/providers/alibaba/SOURCES.md#bailian-function-calling)：fetched |
 | [`glm-4.7`](../../model-sources/providers/alibaba/models/glm-4.7.md) | [`compute/coding-plans/dashscope-token-plan.json`](../../../compute/coding-plans/dashscope-token-plan.json) | `contextWindow`、`maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
 | [`glm-5.2`](../../model-sources/providers/alibaba/models/glm-5.2.md) | [`compute/coding-plans/dashscope-token-plan.json`](../../../compute/coding-plans/dashscope-token-plan.json) | `contextWindow`、`extra.reasoning.defaultEffort`、`extra.reasoning.supportedEfforts`、`maxOutputTokens` | [qwen-models](../../model-sources/providers/alibaba/SOURCES.md#qwen-models)：fetched |
 | [`happyhorse-1.0-i2v`](../../model-sources/providers/alibaba/models/happyhorse-1.0-i2v.md) | [`compute/providers/dashscope.json`](../../../compute/providers/dashscope.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
@@ -71,11 +71,11 @@
 | [`qwen-image-3.0-pro`](../../model-sources/providers/alibaba/models/qwen-image-3.0-pro.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [qwen-models](../../model-sources/providers/alibaba/SOURCES.md#qwen-models)：fetched |
 | [`qwen-long`](../../model-sources/providers/alibaba/models/qwen-long.md) | [`compute/providers/dashscope.json`](../../../compute/providers/dashscope.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`inputPrice`、`maxOutputTokens`、`outputPrice` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
 | [`qwen-long`](../../model-sources/providers/alibaba/models/qwen-long.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
-| [`qwen-max`](../../model-sources/providers/alibaba/models/qwen-max.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
+| [`qwen-max`](../../model-sources/providers/alibaba/models/qwen-max.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched；[bailian-qwen-max-caps](../../model-sources/providers/alibaba/SOURCES.md#bailian-qwen-max-caps)：fetched；[bailian-function-calling](../../model-sources/providers/alibaba/SOURCES.md#bailian-function-calling)：fetched |
 | [`qwen-mt-plus`](../../model-sources/providers/alibaba/models/qwen-mt-plus.md) | [`compute/providers/dashscope.json`](../../../compute/providers/dashscope.json) | `contextWindow`、`inputPrice`、`maxOutputTokens`、`outputPrice` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
 | [`qwen-omni-turbo`](../../model-sources/providers/alibaba/models/qwen-omni-turbo.md) | [`compute/providers/dashscope.json`](../../../compute/providers/dashscope.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
 | [`qwen-omni-turbo`](../../model-sources/providers/alibaba/models/qwen-omni-turbo.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
-| [`qwen-plus`](../../model-sources/providers/alibaba/models/qwen-plus.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
+| [`qwen-plus`](../../model-sources/providers/alibaba/models/qwen-plus.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched；[bailian-qwen-plus-caps](../../model-sources/providers/alibaba/SOURCES.md#bailian-qwen-plus-caps)：fetched；[bailian-function-calling](../../model-sources/providers/alibaba/SOURCES.md#bailian-function-calling)：fetched |
 | [`qwen-turbo`](../../model-sources/providers/alibaba/models/qwen-turbo.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
 | [`qwen3-coder-next`](../../model-sources/providers/alibaba/models/qwen3-coder-next.md) | [`compute/coding-plans/dashscope-coding.json`](../../../compute/coding-plans/dashscope-coding.json) | `contextWindow`、`maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
 | [`qwen3-coder-next`](../../model-sources/providers/alibaba/models/qwen3-coder-next.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [qwen-coding](../../model-sources/providers/alibaba/SOURCES.md#qwen-coding)：fetched |
@@ -85,10 +85,10 @@
 | [`qwen3-max`](../../model-sources/providers/alibaba/models/qwen3-max.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens`、`spec.supportsReasoning` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
 | [`qwen3-rerank`](../../model-sources/providers/alibaba/models/qwen3-rerank.md) | [`compute/providers/dashscope.json`](../../../compute/providers/dashscope.json) | `contextWindow`、`inputPrice` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
 | [`qwen3-rerank`](../../model-sources/providers/alibaba/models/qwen3-rerank.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
-| [`qwen3-vl-flash`](../../model-sources/providers/alibaba/models/qwen3-vl-flash.md) | [`compute/providers/dashscope.json`](../../../compute/providers/dashscope.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`inputPrice`、`maxOutputTokens`、`outputPrice` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
-| [`qwen3-vl-flash`](../../model-sources/providers/alibaba/models/qwen3-vl-flash.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
-| [`qwen3-vl-plus`](../../model-sources/providers/alibaba/models/qwen3-vl-plus.md) | [`compute/providers/dashscope.json`](../../../compute/providers/dashscope.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`inputPrice`、`maxOutputTokens`、`outputPrice` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
-| [`qwen3-vl-plus`](../../model-sources/providers/alibaba/models/qwen3-vl-plus.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
+| [`qwen3-vl-flash`](../../model-sources/providers/alibaba/models/qwen3-vl-flash.md) | [`compute/providers/dashscope.json`](../../../compute/providers/dashscope.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`inputPrice`、`maxOutputTokens`、`outputPrice` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched；[bailian-qwen3-vl-flash-caps](../../model-sources/providers/alibaba/SOURCES.md#bailian-qwen3-vl-flash-caps)：fetched；[bailian-function-calling](../../model-sources/providers/alibaba/SOURCES.md#bailian-function-calling)：fetched |
+| [`qwen3-vl-flash`](../../model-sources/providers/alibaba/models/qwen3-vl-flash.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched；[bailian-qwen3-vl-flash-caps](../../model-sources/providers/alibaba/SOURCES.md#bailian-qwen3-vl-flash-caps)：fetched；[bailian-function-calling](../../model-sources/providers/alibaba/SOURCES.md#bailian-function-calling)：fetched |
+| [`qwen3-vl-plus`](../../model-sources/providers/alibaba/models/qwen3-vl-plus.md) | [`compute/providers/dashscope.json`](../../../compute/providers/dashscope.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`inputPrice`、`maxOutputTokens`、`outputPrice` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched；[bailian-qwen3-vl-plus-caps](../../model-sources/providers/alibaba/SOURCES.md#bailian-qwen3-vl-plus-caps)：fetched；[bailian-function-calling](../../model-sources/providers/alibaba/SOURCES.md#bailian-function-calling)：fetched |
+| [`qwen3-vl-plus`](../../model-sources/providers/alibaba/models/qwen3-vl-plus.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched；[bailian-qwen3-vl-plus-caps](../../model-sources/providers/alibaba/SOURCES.md#bailian-qwen3-vl-plus-caps)：fetched；[bailian-function-calling](../../model-sources/providers/alibaba/SOURCES.md#bailian-function-calling)：fetched |
 | [`qwen3.5-27b`](../../model-sources/providers/alibaba/models/qwen3.5-27b.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
 | [`qwen3.5-35b-a3b`](../../model-sources/providers/alibaba/models/qwen3.5-35b-a3b.md) | [`compute/model-specs/qwen.json`](../../../compute/model-specs/qwen.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
 | [`qwen3.5-plus`](../../model-sources/providers/alibaba/models/qwen3.5-plus.md) | [`compute/coding-plans/dashscope-coding.json`](../../../compute/coding-plans/dashscope-coding.json) | `contextWindow`、`maxOutputTokens` | [qwen-pricing](../../model-sources/providers/alibaba/SOURCES.md#qwen-pricing)：fetched |
--- a/docs/source-gaps/providers/baidu.md
+++ b/docs/source-gaps/providers/baidu.md
@@ -18,8 +18,8 @@
 | --- | --- | --- | --- |
 | [`ernie-5.0-thinking-latest`](../../model-sources/providers/baidu/models/ernie-5.0-thinking-latest.md) | [`compute/providers/baidu.json`](../../../compute/providers/baidu.json) | `contextWindow`、`extra.pricingTiers`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [baidu-models](../../model-sources/providers/baidu/SOURCES.md#baidu-models)：fetched |
 | [`ernie-5.0-thinking-latest`](../../model-sources/providers/baidu/models/ernie-5.0-thinking-latest.md) | [`compute/model-specs/baidu.json`](../../../compute/model-specs/baidu.json) | `id`、`spec.contextWindow`、`spec.maxOutputTokens`、`spec.supportsReasoning` | [baidu-models](../../model-sources/providers/baidu/SOURCES.md#baidu-models)：fetched |
-| [`ernie-x1.1`](../../model-sources/providers/baidu/models/ernie-x1.1.md) | [`compute/providers/baidu.json`](../../../compute/providers/baidu.json) | `contextWindow`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [baidu-models](../../model-sources/providers/baidu/SOURCES.md#baidu-models)：fetched |
-| [`ernie-x1.1`](../../model-sources/providers/baidu/models/ernie-x1.1.md) | [`compute/model-specs/baidu.json`](../../../compute/model-specs/baidu.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens`、`spec.supportsReasoning` | [baidu-models](../../model-sources/providers/baidu/SOURCES.md#baidu-models)：fetched |
+| [`ernie-x1.1`](../../model-sources/providers/baidu/models/ernie-x1.1.md) | [`compute/providers/baidu.json`](../../../compute/providers/baidu.json) | `contextWindow`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [baidu-models](../../model-sources/providers/baidu/SOURCES.md#baidu-models)：fetched；[qianfan-function-calling](../../model-sources/providers/baidu/SOURCES.md#qianfan-function-calling)：fetched |
+| [`ernie-x1.1`](../../model-sources/providers/baidu/models/ernie-x1.1.md) | [`compute/model-specs/baidu.json`](../../../compute/model-specs/baidu.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens`、`spec.supportsReasoning` | [baidu-models](../../model-sources/providers/baidu/SOURCES.md#baidu-models)：fetched；[qianfan-function-calling](../../model-sources/providers/baidu/SOURCES.md#qianfan-function-calling)：fetched |
 | [`minimax-m2.7`](../../model-sources/providers/baidu/models/minimax-m2.7.md) | [`compute/coding-plans/baidu-coding.json`](../../../compute/coding-plans/baidu-coding.json) | `modelName` | [baidu-models](../../model-sources/providers/baidu/SOURCES.md#baidu-models)：fetched |

 ## 仅确认 ID，参数来源未登记（13 条）
@@ -34,10 +34,10 @@
 | [`ernie-4.5-turbo-20260402`](../../model-sources/providers/baidu/models/ernie-4.5-turbo-20260402.md) | [`compute/model-specs/baidu.json`](../../../compute/model-specs/baidu.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [baidu-plan](../../model-sources/providers/baidu/SOURCES.md#baidu-plan)：fetched |
 | [`ernie-5.0`](../../model-sources/providers/baidu/models/ernie-5.0.md) | [`compute/providers/baidu.json`](../../../compute/providers/baidu.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`extra.pricingTiers`、`inputPrice`、`maxOutputTokens`、`outputPrice` | [baidu-models](../../model-sources/providers/baidu/SOURCES.md#baidu-models)：fetched |
 | [`ernie-5.0`](../../model-sources/providers/baidu/models/ernie-5.0.md) | [`compute/model-specs/baidu.json`](../../../compute/model-specs/baidu.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [baidu-models](../../model-sources/providers/baidu/SOURCES.md#baidu-models)：fetched |
-| [`glm-5`](../../model-sources/providers/baidu/models/glm-5.md) | [`compute/coding-plans/baidu-coding.json`](../../../compute/coding-plans/baidu-coding.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [baidu-models](../../model-sources/providers/baidu/SOURCES.md#baidu-models)：fetched |
-| [`kimi-k2.5`](../../model-sources/providers/baidu/models/kimi-k2.5.md) | [`compute/coding-plans/baidu-coding.json`](../../../compute/coding-plans/baidu-coding.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [baidu-plan](../../model-sources/providers/baidu/SOURCES.md#baidu-plan)：fetched |
+| [`glm-5`](../../model-sources/providers/baidu/models/glm-5.md) | [`compute/coding-plans/baidu-coding.json`](../../../compute/coding-plans/baidu-coding.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [baidu-models](../../model-sources/providers/baidu/SOURCES.md#baidu-models)：fetched；[qianfan-function-calling](../../model-sources/providers/baidu/SOURCES.md#qianfan-function-calling)：fetched；[qianfan-coding-plan-tools](../../model-sources/providers/baidu/SOURCES.md#qianfan-coding-plan-tools)：fetched |
+| [`kimi-k2.5`](../../model-sources/providers/baidu/models/kimi-k2.5.md) | [`compute/coding-plans/baidu-coding.json`](../../../compute/coding-plans/baidu-coding.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [baidu-plan](../../model-sources/providers/baidu/SOURCES.md#baidu-plan)：fetched；[qianfan-function-calling](../../model-sources/providers/baidu/SOURCES.md#qianfan-function-calling)：fetched；[qianfan-coding-plan-tools](../../model-sources/providers/baidu/SOURCES.md#qianfan-coding-plan-tools)：fetched |
 | [`kimi-k2.6`](../../model-sources/providers/baidu/models/kimi-k2.6.md) | [`compute/coding-plans/baidu-coding.json`](../../../compute/coding-plans/baidu-coding.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [baidu-models](../../model-sources/providers/baidu/SOURCES.md#baidu-models)：fetched |
-| [`minimax-m2.5`](../../model-sources/providers/baidu/models/minimax-m2.5.md) | [`compute/coding-plans/baidu-coding.json`](../../../compute/coding-plans/baidu-coding.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [baidu-plan](../../model-sources/providers/baidu/SOURCES.md#baidu-plan)：fetched |
+| [`minimax-m2.5`](../../model-sources/providers/baidu/models/minimax-m2.5.md) | [`compute/coding-plans/baidu-coding.json`](../../../compute/coding-plans/baidu-coding.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [baidu-plan](../../model-sources/providers/baidu/SOURCES.md#baidu-plan)：fetched；[qianfan-function-calling](../../model-sources/providers/baidu/SOURCES.md#qianfan-function-calling)：fetched；[qianfan-coding-plan-tools](../../model-sources/providers/baidu/SOURCES.md#qianfan-coding-plan-tools)：fetched |
 | [`qianfan-code-latest`](../../model-sources/providers/baidu/models/qianfan-code-latest.md) | [`compute/coding-plans/baidu-coding.json`](../../../compute/coding-plans/baidu-coding.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [baidu-plan](../../model-sources/providers/baidu/SOURCES.md#baidu-plan)：fetched |

 ## 已有部分参数证明，仍需区分未证明字段（2 条）
--- a/docs/source-gaps/providers/ollama.md
+++ b/docs/source-gaps/providers/ollama.md
@@ -16,7 +16,7 @@

 | 模型 | 接入面 | 未登记证明的关键字段 | 官网入口/读取结果 |
 | --- | --- | --- | --- |
-| [`llama3.1:70b`](../../model-sources/providers/ollama/models/llama3.1--70b.md) | [`compute/providers/ollama.json`](../../../compute/providers/ollama.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`maxOutputTokens`、`modelName` | [ollama](../../model-sources/providers/ollama/SOURCES.md#ollama)：fetched |
+| [`llama3.1:70b`](../../model-sources/providers/ollama/models/llama3.1--70b.md) | [`compute/providers/ollama.json`](../../../compute/providers/ollama.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`maxOutputTokens`、`modelName` | [ollama](../../model-sources/providers/ollama/SOURCES.md#ollama)：fetched；[ollama-llama31-70b](../../model-sources/providers/ollama/SOURCES.md#ollama-llama31-70b)：fetched |

 ## 仅确认 ID，参数来源未登记（0 条）

--- a/docs/source-gaps/providers/openai.md
+++ b/docs/source-gaps/providers/openai.md
@@ -19,8 +19,8 @@
 | [`dall-e-3`](../../model-sources/providers/openai/models/dall-e-3.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
 | [`gpt-4.1-mini`](../../model-sources/providers/openai/models/gpt-4.1-mini.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`extra.serverSideWebSearch.searchRequestPriceUsd`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
 | [`gpt-4.1-mini`](../../model-sources/providers/openai/models/gpt-4.1-mini.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
-| [`gpt-4.1-nano`](../../model-sources/providers/openai/models/gpt-4.1-nano.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
-| [`gpt-4.1-nano`](../../model-sources/providers/openai/models/gpt-4.1-nano.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
+| [`gpt-4.1-nano`](../../model-sources/providers/openai/models/gpt-4.1-nano.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched；[openai-gpt-4-1-nano-tools](../../model-sources/providers/openai/SOURCES.md#openai-gpt-4-1-nano-tools)：fetched |
+| [`gpt-4.1-nano`](../../model-sources/providers/openai/models/gpt-4.1-nano.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched；[openai-gpt-4-1-nano-tools](../../model-sources/providers/openai/SOURCES.md#openai-gpt-4-1-nano-tools)：fetched |
 | [`gpt-4.1`](../../model-sources/providers/openai/models/gpt-4.1.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`extra.serverSideWebSearch.searchRequestPriceUsd`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
 | [`gpt-4.1`](../../model-sources/providers/openai/models/gpt-4.1.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
 | [`gpt-4o-mini`](../../model-sources/providers/openai/models/gpt-4o-mini.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
@@ -29,8 +29,8 @@
 | [`gpt-4o-realtime`](../../model-sources/providers/openai/models/gpt-4o-realtime.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
 | [`gpt-5-mini`](../../model-sources/providers/openai/models/gpt-5-mini.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `contextWindow`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
 | [`gpt-5-mini`](../../model-sources/providers/openai/models/gpt-5-mini.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens`、`spec.supportsReasoning` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
-| [`gpt-5-nano`](../../model-sources/providers/openai/models/gpt-5-nano.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `contextWindow`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
-| [`gpt-5-nano`](../../model-sources/providers/openai/models/gpt-5-nano.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
+| [`gpt-5-nano`](../../model-sources/providers/openai/models/gpt-5-nano.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `contextWindow`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched；[openai-gpt-5-nano-tools](../../model-sources/providers/openai/SOURCES.md#openai-gpt-5-nano-tools)：fetched |
+| [`gpt-5-nano`](../../model-sources/providers/openai/models/gpt-5-nano.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched；[openai-gpt-5-nano-tools](../../model-sources/providers/openai/SOURCES.md#openai-gpt-5-nano-tools)：fetched |
 | [`gpt-5-pro`](../../model-sources/providers/openai/models/gpt-5-pro.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `contextWindow`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
 | [`gpt-5-pro`](../../model-sources/providers/openai/models/gpt-5-pro.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens`、`spec.supportsReasoning` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
 | [`gpt-5.1`](../../model-sources/providers/openai/models/gpt-5.1.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `contextWindow`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
@@ -64,7 +64,7 @@
 | [`gpt-image-2.5-sunburst`](../../model-sources/providers/openai/models/gpt-image-2.5-sunburst.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
 | [`gpt-image-2`](../../model-sources/providers/openai/models/gpt-image-2.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `extra.imageCachedInputPrice`、`extra.imageInputPrice`、`extra.textCachedInputPrice`、`inputPrice`、`modelName`、`outputPrice` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
 | [`gpt-image-2`](../../model-sources/providers/openai/models/gpt-image-2.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
-| [`gpt-oss-120b`](../../model-sources/providers/openai/models/gpt-oss-120b.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow`、`spec.maxOutputTokens` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
+| [`gpt-oss-120b`](../../model-sources/providers/openai/models/gpt-oss-120b.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow`、`spec.maxOutputTokens` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched；[openai-gpt-oss-120b-tools](../../model-sources/providers/openai/SOURCES.md#openai-gpt-oss-120b-tools)：fetched |
 | [`gpt-realtime-2.1-mini`](../../model-sources/providers/openai/models/gpt-realtime-2.1-mini.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `contextWindow`、`extra.audioCachedInputPrice`、`extra.audioInputPrice`、`extra.audioOutputPrice`、`extra.imageCachedInputPrice`、`extra.imageInputPrice`、`extra.textCachedInputPrice`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
 | [`gpt-reserve`](../../model-sources/providers/openai/models/gpt-reserve.md) | [`compute/providers/openai-codex.json`](../../../compute/providers/openai-codex.json) | `contextWindow`、`extra.reasoning.defaultEffort`、`extra.reasoning.supportedEfforts`、`maxOutputTokens`、`modelName` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
 | [`gpt-reserve`](../../model-sources/providers/openai/models/gpt-reserve.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `id`、`spec.contextWindow`、`spec.maxOutputTokens`、`spec.supportsReasoning` | [openai-models](../../model-sources/providers/openai/SOURCES.md#openai-models)：fetched |
@@ -109,7 +109,7 @@
 | --- | --- | --- | --- |
 | [`gpt-5.4-mini`](../../model-sources/providers/openai/models/gpt-5.4-mini.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `extra.serverSideWebSearch.searchRequestPriceUsd` | [gpt-5-4-mini-api](../../model-sources/providers/openai/SOURCES.md#gpt-5-4-mini-api)：fetched |
 | [`gpt-5.4-mini`](../../model-sources/providers/openai/models/gpt-5.4-mini.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | `spec.supportsReasoning` | [gpt-5-4-mini-api](../../model-sources/providers/openai/SOURCES.md#gpt-5-4-mini-api)：fetched |
-| [`gpt-5.4-nano`](../../model-sources/providers/openai/models/gpt-5.4-nano.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `extra.serverSideWebSearch.searchRequestPriceUsd` | [gpt-5-4-nano-api](../../model-sources/providers/openai/SOURCES.md#gpt-5-4-nano-api)：fetched |
+| [`gpt-5.4-nano`](../../model-sources/providers/openai/models/gpt-5.4-nano.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `extra.serverSideWebSearch.searchRequestPriceUsd` | [gpt-5-4-nano-api](../../model-sources/providers/openai/SOURCES.md#gpt-5-4-nano-api)：fetched；[openai-gpt-5-4-nano-tools](../../model-sources/providers/openai/SOURCES.md#openai-gpt-5-4-nano-tools)：fetched |
 | [`gpt-5.4-nano`](../../model-sources/providers/openai/models/gpt-5.4-nano.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [gpt-5-4-nano-api](../../model-sources/providers/openai/SOURCES.md#gpt-5-4-nano-api)：fetched |
 | [`gpt-5.4-pro`](../../model-sources/providers/openai/models/gpt-5.4-pro.md) | [`compute/providers/openai.json`](../../../compute/providers/openai.json) | `extra.serverSideWebSearch.searchRequestPriceUsd` | [gpt-5-4-pro-api](../../model-sources/providers/openai/SOURCES.md#gpt-5-4-pro-api)：fetched |
 | [`gpt-5.4-pro`](../../model-sources/providers/openai/models/gpt-5.4-pro.md) | [`compute/model-specs/openai.json`](../../../compute/model-specs/openai.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [gpt-5-4-pro-api](../../model-sources/providers/openai/SOURCES.md#gpt-5-4-pro-api)：fetched |
--- a/docs/source-gaps/providers/openrouter.md
+++ b/docs/source-gaps/providers/openrouter.md
@@ -33,7 +33,7 @@
 | [`mistralai/mistral-medium-3-5`](../../model-sources/providers/openrouter/models/mistralai--mistral-medium-3-5.md) | [`compute/providers/openrouter.json`](../../../compute/providers/openrouter.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [openrouter-api](../../model-sources/providers/openrouter/SOURCES.md#openrouter-api)：fetched |
 | [`moonshotai/kimi-k3`](../../model-sources/providers/openrouter/models/moonshotai--kimi-k3.md) | [`compute/providers/openrouter.json`](../../../compute/providers/openrouter.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [openrouter-api](../../model-sources/providers/openrouter/SOURCES.md#openrouter-api)：fetched |
 | [`openai/gpt-6.1-sol`](../../model-sources/providers/openrouter/models/openai--gpt-6.1-sol.md) | [`compute/providers/openrouter.json`](../../../compute/providers/openrouter.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [openrouter-api](../../model-sources/providers/openrouter/SOURCES.md#openrouter-api)：fetched |
-| [`openrouter/auto`](../../model-sources/providers/openrouter/models/openrouter--auto.md) | [`compute/providers/openrouter.json`](../../../compute/providers/openrouter.json) | `defaultTemperature`、`defaultTopP`、`maxOutputTokens` | [openrouter-api](../../model-sources/providers/openrouter/SOURCES.md#openrouter-api)：fetched |
+| [`openrouter/auto`](../../model-sources/providers/openrouter/models/openrouter--auto.md) | [`compute/providers/openrouter.json`](../../../compute/providers/openrouter.json) | `defaultTemperature`、`defaultTopP`、`maxOutputTokens` | [openrouter-api](../../model-sources/providers/openrouter/SOURCES.md#openrouter-api)：fetched；[openrouter-models-tools](../../model-sources/providers/openrouter/SOURCES.md#openrouter-models-tools)：fetched |
 | [`qwen/qwen3.8-max-prime`](../../model-sources/providers/openrouter/models/qwen--qwen3.8-max-prime.md) | [`compute/providers/openrouter.json`](../../../compute/providers/openrouter.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [openrouter-api](../../model-sources/providers/openrouter/SOURCES.md#openrouter-api)：fetched |
 | [`qwen/qwen3.8-omni-flash`](../../model-sources/providers/openrouter/models/qwen--qwen3.8-omni-flash.md) | [`compute/providers/openrouter.json`](../../../compute/providers/openrouter.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [openrouter-api](../../model-sources/providers/openrouter/SOURCES.md#openrouter-api)：fetched |
 | [`tencent/hy4-preview`](../../model-sources/providers/openrouter/models/tencent--hy4-preview.md) | [`compute/providers/openrouter.json`](../../../compute/providers/openrouter.json) | 已登记关键字段都有证明；其他合同限制仍按原记录复核 | [openrouter-api](../../model-sources/providers/openrouter/SOURCES.md#openrouter-api)：fetched |
--- a/docs/source-gaps/providers/tencent.md
+++ b/docs/source-gaps/providers/tencent.md
@@ -37,8 +37,8 @@
 | [`hy3-preview`](../../model-sources/providers/tencent/models/hy3-preview.md) | [`compute/model-specs/tencent.json`](../../../compute/model-specs/tencent.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.supportsReasoning` | [tencent](../../model-sources/providers/tencent/SOURCES.md#tencent)：fetched |
 | [`kimi-k2.5`](../../model-sources/providers/tencent/models/kimi-k2.5.md) | [`compute/coding-plans/tencent-token.json`](../../../compute/coding-plans/tencent-token.json) | `modelName` | [tencent](../../model-sources/providers/tencent/SOURCES.md#tencent)：fetched |
 | [`minimax-m2.5`](../../model-sources/providers/tencent/models/minimax-m2.5.md) | [`compute/coding-plans/tencent-token.json`](../../../compute/coding-plans/tencent-token.json) | `modelName` | [tencent](../../model-sources/providers/tencent/SOURCES.md#tencent)：fetched |
-| [`minimax-m2.7`](../../model-sources/providers/tencent/models/minimax-m2.7.md) | [`compute/coding-plans/tencent-token.json`](../../../compute/coding-plans/tencent-token.json) | `modelName` | [tencent](../../model-sources/providers/tencent/SOURCES.md#tencent)：fetched |
-| [`tc-code-latest`](../../model-sources/providers/tencent/models/tc-code-latest.md) | [`compute/coding-plans/tencent-token.json`](../../../compute/coding-plans/tencent-token.json) | `modelName` | [tencent](../../model-sources/providers/tencent/SOURCES.md#tencent)：fetched |
+| [`minimax-m2.7`](../../model-sources/providers/tencent/models/minimax-m2.7.md) | [`compute/coding-plans/tencent-token.json`](../../../compute/coding-plans/tencent-token.json) | `modelName` | [tencent](../../model-sources/providers/tencent/SOURCES.md#tencent)：fetched；[tokenhub-model-list](../../model-sources/providers/tencent/SOURCES.md#tokenhub-model-list)：fetched；[tencent-token-plan-doc](../../model-sources/providers/tencent/SOURCES.md#tencent-token-plan-doc)：fetched |
+| [`tc-code-latest`](../../model-sources/providers/tencent/models/tc-code-latest.md) | [`compute/coding-plans/tencent-token.json`](../../../compute/coding-plans/tencent-token.json) | `modelName` | [tencent](../../model-sources/providers/tencent/SOURCES.md#tencent)：fetched；[tencent-token-plan-doc](../../model-sources/providers/tencent/SOURCES.md#tencent-token-plan-doc)：fetched；[tokenhub-model-list](../../model-sources/providers/tencent/SOURCES.md#tokenhub-model-list)：fetched |

 ## 仅确认 ID，参数来源未登记（1 条）

--- a/docs/source-gaps/providers/volcengine.md
+++ b/docs/source-gaps/providers/volcengine.md
@@ -8,11 +8,10 @@

 无该类记录。

-## 未取得可读官网证据（66 条）
+## 未取得可读官网证据（65 条）

 | 模型 | 接入面 | 未登记证明的关键字段 | 官网入口/读取结果 |
 | --- | --- | --- | --- |
-| [`ark-code-latest`](../../model-sources/providers/volcengine/models/ark-code-latest.md) | [`compute/coding-plans/volcengine-coding.json`](../../../compute/coding-plans/volcengine-coding.json) | `modelName` | [volcengine-models](../../model-sources/providers/volcengine/SOURCES.md#volcengine-models)：shell |
 | [`deepseek-r1`](../../model-sources/providers/volcengine/models/deepseek-r1.md) | [`compute/providers/volcengine.json`](../../../compute/providers/volcengine.json) | `contextWindow`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [volcengine-models](../../model-sources/providers/volcengine/SOURCES.md#volcengine-models)：shell |
 | [`deepseek-r1`](../../model-sources/providers/volcengine/models/deepseek-r1.md) | [`compute/model-specs/volcengine.json`](../../../compute/model-specs/volcengine.json) | `id`、`spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens`、`spec.supportsReasoning` | [volcengine-models](../../model-sources/providers/volcengine/SOURCES.md#volcengine-models)：shell |
 | [`deepseek-v3.2`](../../model-sources/providers/volcengine/models/deepseek-v3.2.md) | [`compute/providers/volcengine.json`](../../../compute/providers/volcengine.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`inputPrice`、`maxOutputTokens`、`modelName`、`outputPrice` | [volcengine-models](../../model-sources/providers/volcengine/SOURCES.md#volcengine-models)：shell |
@@ -79,9 +78,11 @@
 | [`volc-translation`](../../model-sources/providers/volcengine/models/volc-translation.md) | [`compute/providers/volcengine.json`](../../../compute/providers/volcengine.json) | `modelName` | [volcengine-models](../../model-sources/providers/volcengine/SOURCES.md#volcengine-models)：shell |
 | [`volc-translation`](../../model-sources/providers/volcengine/models/volc-translation.md) | [`compute/model-specs/volcengine.json`](../../../compute/model-specs/volcengine.json) | `id` | [volcengine-models](../../model-sources/providers/volcengine/SOURCES.md#volcengine-models)：shell |

-## 入口有读取记录，但本条模型无字段证明（0 条）
+## 入口有读取记录，但本条模型无字段证明（1 条）

-无该类记录。
+| 模型 | 接入面 | 未登记证明的关键字段 | 官网入口/读取结果 |
+| --- | --- | --- | --- |
+| [`ark-code-latest`](../../model-sources/providers/volcengine/models/ark-code-latest.md) | [`compute/coding-plans/volcengine-coding.json`](../../../compute/coding-plans/volcengine-coding.json) | `modelName` | [volcengine-models](../../model-sources/providers/volcengine/SOURCES.md#volcengine-models)：shell；[volcengine-coding-get-started](../../model-sources/providers/volcengine/SOURCES.md#volcengine-coding-get-started)：fetched；[volcengine-coding-overview](../../model-sources/providers/volcengine/SOURCES.md#volcengine-coding-overview)：fetched；[volcengine-models-rendered](../../model-sources/providers/volcengine/SOURCES.md#volcengine-models-rendered)：fetched |

 ## 仅确认 ID，参数来源未登记（0 条）

--- a/docs/source-gaps/providers/zhipu.md
+++ b/docs/source-gaps/providers/zhipu.md
@@ -27,10 +27,10 @@
 | [`embedding-3`](../../model-sources/providers/zhipu/models/embedding-3.md) | [`compute/providers/zhipu-embedding.json`](../../../compute/providers/zhipu-embedding.json) | `contextWindow`、`inputPrice` | [glm-pricing](../../model-sources/providers/zhipu/SOURCES.md#glm-pricing)：fetched |
 | [`embedding-3`](../../model-sources/providers/zhipu/models/embedding-3.md) | [`compute/model-specs/zhipu.json`](../../../compute/model-specs/zhipu.json) | `spec.contextWindow` | [glm-pricing](../../model-sources/providers/zhipu/SOURCES.md#glm-pricing)：fetched |
 | [`glm-4.6`](../../model-sources/providers/zhipu/models/glm-4.6.md) | [`compute/providers/zhipu.json`](../../../compute/providers/zhipu.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`extra.cacheHitPrice`、`inputPrice`、`maxOutputTokens`、`outputPrice` | [glm-pricing](../../model-sources/providers/zhipu/SOURCES.md#glm-pricing)：fetched |
-| [`glm-4.6`](../../model-sources/providers/zhipu/models/glm-4.6.md) | [`compute/model-specs/zhipu.json`](../../../compute/model-specs/zhipu.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens`、`spec.supportsReasoning` | [glm-pricing](../../model-sources/providers/zhipu/SOURCES.md#glm-pricing)：fetched |
+| [`glm-4.6`](../../model-sources/providers/zhipu/models/glm-4.6.md) | [`compute/model-specs/zhipu.json`](../../../compute/model-specs/zhipu.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens`、`spec.supportsReasoning` | [glm-pricing](../../model-sources/providers/zhipu/SOURCES.md#glm-pricing)：fetched；[zhipu-glm-4-6-page](../../model-sources/providers/zhipu/SOURCES.md#zhipu-glm-4-6-page)：fetched |
 | [`glm-4.6v`](../../model-sources/providers/zhipu/models/glm-4.6v.md) | [`compute/providers/zhipu.json`](../../../compute/providers/zhipu.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`extra.cacheHitPrice`、`extra.cacheStoragePrice`、`extra.pricingTiers`、`inputPrice`、`maxOutputTokens`、`outputPrice` | [glm-pricing](../../model-sources/providers/zhipu/SOURCES.md#glm-pricing)：fetched |
 | [`glm-4.6v`](../../model-sources/providers/zhipu/models/glm-4.6v.md) | [`compute/model-specs/zhipu.json`](../../../compute/model-specs/zhipu.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens` | [glm-pricing](../../model-sources/providers/zhipu/SOURCES.md#glm-pricing)：fetched |
-| [`glm-4.7-flashx`](../../model-sources/providers/zhipu/models/glm-4.7-flashx.md) | [`compute/providers/zhipu.json`](../../../compute/providers/zhipu.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`extra.cacheHitPrice`、`extra.cacheStoragePrice`、`inputPrice`、`maxOutputTokens`、`outputPrice` | [glm-pricing](../../model-sources/providers/zhipu/SOURCES.md#glm-pricing)：fetched |
+| [`glm-4.7-flashx`](../../model-sources/providers/zhipu/models/glm-4.7-flashx.md) | [`compute/providers/zhipu.json`](../../../compute/providers/zhipu.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`extra.cacheHitPrice`、`extra.cacheStoragePrice`、`inputPrice`、`maxOutputTokens`、`outputPrice` | [glm-pricing](../../model-sources/providers/zhipu/SOURCES.md#glm-pricing)：fetched；[zhipu-glm-4-7-page](../../model-sources/providers/zhipu/SOURCES.md#zhipu-glm-4-7-page)：fetched；[zhipu-model-overview](../../model-sources/providers/zhipu/SOURCES.md#zhipu-model-overview)：fetched |
 | [`glm-4.7`](../../model-sources/providers/zhipu/models/glm-4.7.md) | [`compute/providers/zhipu.json`](../../../compute/providers/zhipu.json) | `contextWindow`、`defaultTemperature`、`defaultTopP`、`extra.cacheHitPrice`、`extra.cacheStoragePrice`、`extra.pricingTiers`、`inputPrice`、`maxOutputTokens`、`outputPrice` | [glm-pricing](../../model-sources/providers/zhipu/SOURCES.md#glm-pricing)：fetched |
 | [`glm-4.7`](../../model-sources/providers/zhipu/models/glm-4.7.md) | [`compute/coding-plans/zhipu-coding.json`](../../../compute/coding-plans/zhipu-coding.json) | `contextWindow`、`maxOutputTokens` | [glm-pricing](../../model-sources/providers/zhipu/SOURCES.md#glm-pricing)：fetched |
 | [`glm-4.7`](../../model-sources/providers/zhipu/models/glm-4.7.md) | [`compute/model-specs/zhipu.json`](../../../compute/model-specs/zhipu.json) | `spec.contextWindow`、`spec.defaultTemperature`、`spec.maxOutputTokens`、`spec.supportsReasoning` | [glm-pricing](../../model-sources/providers/zhipu/SOURCES.md#glm-pricing)：fetched |
--- a/docs/source-rechecks/2026-10-05-tool-calling/README.md
+++ b/docs/source-rechecks/2026-10-05-tool-calling/README.md
@@ -0,0 +1,174 @@
+# 2026-10-05：工具调用标签核对
+
+基线 main `a94a751078f81a69436e05e29b751dfea163f685`，数据版本 128 → 129。
+
+范围：`capabilities` 含 `chat`、但没有 `tool_use` / `tools` / `function_calling` 的 36 个共享规格与 70 个 Provider / 套餐预置条目（同一型号在不同接入面分别计数）。逐条读取供应商官方文档，分地域、分接入面判断；没有调用任何模型，也没有做账号调用验收。
+
+结果：25 条补 `tool_use`（规格 9、随规格继承的接入条目 5、接入覆盖 11），81 条保持原样。只改 `capabilities`，没有改 `routing`；新补标签的规格都没有路由策略，自动路由范围不变。
+
+## 判据
+
+| 情况 | 处理 |
+| --- | --- |
+| 官方页面写明该型号（或所属系列）在该接入面支持函数 / 工具调用 | 补 `tool_use` |
+| 套餐别名：官方套餐文档把它作为 Claude Code 等编程工具里配置的模型名，且套餐内可选型号在同平台文档里都标注支持 | 补 `tool_use`，依据类型记为套餐文档 |
+| 官方写明不支持、官方页面互相矛盾、官方当前文档里没有该型号 | 不改 |
+| 官方已下线或公告数日内下线 | 不改 |
+| 官网读取失败 | 不改 |
+
+来源只取各供应商官方域名上成功读取的正文。能力标签里只有 `tool_use` 一项以这些来源为依据，其余标签仍是本仓产品标签；各记录的核验状态没有因此提高。
+
+## 已补 `tool_use`（25 条）
+
+| 配置 | 型号 | 依据 |
+| --- | --- | --- |
+| `model-specs/openai.json` | `gpt-5-nano` | 官网型号页 Supported features 列有 function_calling，Chat Completions 与 Responses 端点均为 Supported。来源：[developers.openai.com](https://developers.openai.com/api/docs/models/gpt-5-nano.md) |
+| `model-specs/openai.json` | `gpt-4.1-nano` | 官网型号页 Supported features 列有 function_calling，Chat Completions 与 Responses 端点均为 Supported。来源：[developers.openai.com](https://developers.openai.com/api/docs/models/gpt-4.1-nano.md) |
+| `model-specs/openai.json` | `gpt-oss-120b` | 官网型号页 Supported features 列有 function_calling；OpenAI 直连只有 Responses 端点为 Supported（Chat Completions 为 Not supported），其他托管平台另核。来源：[developers.openai.com](https://developers.openai.com/api/docs/models/gpt-oss-120b.md) |
+| `model-specs/qwen.json` | `qwen-max` | 百炼型号页模型能力表：华北 2（北京）Function Calling 支持，新加坡（国际）不支持；Function Calling 指南的支持清单列有 Qwen-Max 系列。标签按北京地域记，国际站接入不适用。来源：[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-max)、[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-function-calling) |
+| `model-specs/qwen.json` | `qwen-plus` | 百炼型号页模型能力表：华北 2（北京）Function Calling 支持，新加坡、法兰克福、弗吉尼亚、中国香港不支持；Function Calling 指南的支持清单列有 Qwen-Plus 系列。标签按北京地域记，其他地域接入不适用。来源：[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-plus)、[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-function-calling) |
+| `model-specs/qwen.json` | `qwen3-vl-plus` | 百炼型号页模型能力表：华北 2（北京）Function Calling 支持，新加坡、法兰克福、弗吉尼亚、中国香港不支持；Function Calling 指南的支持清单列有 Qwen3-VL-Plus 系列。标签按北京地域记，其他地域接入不适用。来源：[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen3-vl-plus)、[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-function-calling) |
+| `model-specs/qwen.json` | `qwen3-vl-flash` | 百炼型号页模型能力表：华北 2（北京）Function Calling 支持，新加坡、法兰克福、弗吉尼亚不支持；Function Calling 指南的支持清单列有 Qwen3-VL-Flash 系列。标签按北京地域记，其他地域接入不适用。来源：[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen3-vl-flash)、[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-function-calling) |
+| `model-specs/zhipu.json` | `glm-4.6` | 官网 GLM-4.6 型号页「能力支持」列有 Function Calling。来源：[docs.bigmodel.cn](https://docs.bigmodel.cn/cn/guide/models/text/glm-4.6.md) |
+| `model-specs/baidu.json` | `ernie-x1.1` | 千帆 Function calling 文档「支持模型范围」的 ERNIE 系列列有 ernie-x1.1。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5) |
+| `providers/openai.json` | `gpt-5-nano` | 随共享规格继承。官网型号页 Supported features 列有 function_calling，Chat Completions 端点为 Supported。来源：[developers.openai.com](https://developers.openai.com/api/docs/models/gpt-5-nano.md) |
+| `providers/openai.json` | `gpt-4.1-nano` | 随共享规格继承。官网型号页 Supported features 列有 function_calling，Chat Completions 端点为 Supported。来源：[developers.openai.com](https://developers.openai.com/api/docs/models/gpt-4.1-nano.md) |
+| `providers/dashscope.json` | `qwen3-vl-plus` | 随共享规格继承。本接入是华北 2（北京）兼容端点，型号页北京地域 Function Calling 支持。来源：[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen3-vl-plus)、[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-function-calling) |
+| `providers/dashscope.json` | `qwen3-vl-flash` | 随共享规格继承。本接入是华北 2（北京）兼容端点，型号页北京地域 Function Calling 支持。来源：[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen3-vl-flash)、[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-function-calling) |
+| `providers/baidu.json` | `ernie-x1.1` | 随共享规格继承。千帆 Function calling 文档「支持模型范围」的 ERNIE 系列列有 ernie-x1.1。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5) |
+| `providers/openai.json` | `gpt-5.4-nano` | 官网型号页 Supported features 列有 function_calling，Chat Completions 与 Responses 端点均为 Supported。来源：[developers.openai.com](https://developers.openai.com/api/docs/models/gpt-5.4-nano.md) |
+| `providers/zhipu.json` | `glm-4.7-flashx` | 官网 GLM-4.7 系列页含 GLM-4.7-FlashX 页签，系列「能力支持」列有 Function Calling；模型概览的 GLM-4.7-FlashX 条目指向该页。精确 API ID 仍未单独证实，绑定方式不变。来源：[docs.bigmodel.cn](https://docs.bigmodel.cn/cn/guide/models/text/glm-4.7.md)、[docs.bigmodel.cn](https://docs.bigmodel.cn/cn/guide/start/model-overview.md) |
+| `providers/ollama.json` | `llama3.1:70b` | Ollama 官方模型库 llama3.1:70b 页带 tools 能力标记。本地实际能力取决于用户拉取的权重与 Ollama 版本。来源：[ollama.com](https://ollama.com/library/llama3.1:70b) |
+| `providers/openrouter.json` | `openrouter/auto` | OpenRouter 官方模型目录 API 中 openrouter/auto 的 supported_parameters 含 tools、tool_choice。实际能力取决于被路由到的后端模型。来源：[openrouter.ai](https://openrouter.ai/api/v1/models) |
+| `coding-plans/dashscope-coding.json` | `glm-4.7` | 百炼 glm-4.7 型号页模型能力表 Function Calling 支持；Coding Plan 概述的支持模型含 glm-4.7。百炼 Function Calling 指南注明调用 GLM 系列需在请求中传 tool_stream=true，否则不返回 tool_calls；套餐端点是否同样要求未核。来源：[help.aliyun.com](https://help.aliyun.com/zh/model-studio/glm-4-7)、[help.aliyun.com](https://help.aliyun.com/zh/model-studio/coding-plan)、[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-function-calling) |
+| `coding-plans/volcengine-coding.json` | `ark-code-latest` | Coding Plan 快速开始把 ark-code-latest 作为 Claude Code 的 ANTHROPIC_MODEL 配置；套餐概览写明套餐额度仅在 AI 编程工具中生效；方舟模型列表「工具调用能力」表里套餐内的豆包、GLM、DeepSeek 型号函数调用均为支持。别名本身没有单独的能力表，依据是套餐文档。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-get-started?lang=zh)、[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-plan-overview?lang=zh)、[docs.volcengine.com](https://docs.volcengine.com/docs/ark/model-list?lang=zh) |
+| `coding-plans/tencent-token.json` | `tc-code-latest` | Token Plan 文档把 tc-code-latest 列为 Auto 智能路由的 Model ID，套餐适配 Claude Code、OpenCode 等编程工具；TokenHub 模型列表里套餐当前可用型号的能力支持均含 Function Calling。别名本身没有单独的能力表，依据是套餐文档。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130060)、[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130051) |
+| `coding-plans/tencent-token.json` | `minimax-m2.7` | TokenHub 模型列表 MiniMax-M2.7（minimax-m2.7）的能力支持含 Function Calling；Token Plan 文档的可用模型含 minimax-m2.7。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130051)、[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130060) |
+| `coding-plans/baidu-coding.json` | `kimi-k2.5` | 千帆 Function calling 文档「支持模型范围」列有 kimi-k2.5；Coding Plan 文档可配置的 Model Name 含 kimi-k2.5。套餐当前可用性未重新核实。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5)、[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan/s/imlg0beiu) |
+| `coding-plans/baidu-coding.json` | `glm-5` | 千帆 Function calling 文档「支持模型范围」列有 glm-5；Coding Plan 文档可配置的 Model Name 含 glm-5。套餐当前可用性未重新核实。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5)、[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan/s/imlg0beiu) |
+| `coding-plans/baidu-coding.json` | `minimax-m2.5` | 千帆 Function calling 文档「支持模型范围」列有 minimax-m2.5；Coding Plan 文档可配置的 Model Name 含 minimax-m2.5（套餐表标注即将下线）。套餐当前可用性未重新核实。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5)、[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan/s/imlg0beiu) |
+
+## 保持原样（81 条）
+
+### 官方写明不支持，或官方页面互相矛盾（10 条）
+
+| 配置 | 型号 | 原因 |
+| --- | --- | --- |
+| `model-specs/qwen.json` | `qwen-turbo` | 型号页模型能力表里华北 2（北京）与新加坡都写 Function Calling 不支持；Function Calling 指南的支持清单又列有 Qwen-Turbo 系列。两份官方页面不一致，不改。来源：[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-turbo)、[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-function-calling) |
+| `model-specs/qwen.json` | `qwen-long` | 型号页模型能力表写 Function Calling 不支持；Function Calling 指南的支持清单也没有 Qwen-Long。来源：[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-long)、[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-function-calling) |
+| `model-specs/perplexity.json` | `sonar-pro` | 官网「Sonar vs Agent API」对比里，自定义函数与 MCP 只属于 Agent API；Sonar Chat Completions 的支持已于 2026-09-27 结束。来源：[docs.perplexity.ai](https://docs.perplexity.ai/docs/agent-api/migrate-from-sonar/overview.md) |
+| `model-specs/perplexity.json` | `sonar-reasoning-pro` | 官网「Sonar vs Agent API」对比里，自定义函数与 MCP 只属于 Agent API；Sonar Chat Completions 的支持已于 2026-09-27 结束。来源：[docs.perplexity.ai](https://docs.perplexity.ai/docs/agent-api/migrate-from-sonar/overview.md) |
+| `model-specs/perplexity.json` | `sonar` | 官网「Sonar vs Agent API」对比里，自定义函数与 MCP 只属于 Agent API；Sonar Chat Completions 的支持已于 2026-09-27 结束。来源：[docs.perplexity.ai](https://docs.perplexity.ai/docs/agent-api/migrate-from-sonar/overview.md) |
+| `providers/dashscope.json` | `qwen-long` | 型号页模型能力表写 Function Calling 不支持；Function Calling 指南的支持清单也没有 Qwen-Long。来源：[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-long)、[help.aliyun.com](https://help.aliyun.com/zh/model-studio/qwen-function-calling) |
+| `providers/perplexity.json` | `sonar-pro` | 官网「Sonar vs Agent API」对比里，自定义函数与 MCP 只属于 Agent API；Sonar Chat Completions 的支持已于 2026-09-27 结束。来源：[docs.perplexity.ai](https://docs.perplexity.ai/docs/agent-api/migrate-from-sonar/overview.md) |
+| `providers/perplexity.json` | `sonar-reasoning-pro` | 官网「Sonar vs Agent API」对比里，自定义函数与 MCP 只属于 Agent API；Sonar Chat Completions 的支持已于 2026-09-27 结束。来源：[docs.perplexity.ai](https://docs.perplexity.ai/docs/agent-api/migrate-from-sonar/overview.md) |
+| `providers/perplexity.json` | `sonar` | 官网「Sonar vs Agent API」对比里，自定义函数与 MCP 只属于 Agent API；Sonar Chat Completions 的支持已于 2026-09-27 结束。来源：[docs.perplexity.ai](https://docs.perplexity.ai/docs/agent-api/migrate-from-sonar/overview.md) |
+| `coding-plans/baidu-coding.json` | `qianfan-code-latest` | 别名由控制台选择实际型号；可选型号里 ernie-4.5-turbo-20260402、deepseek-v4-flash、glm-5.1 不在千帆 Function calling 的支持模型范围内，依据不一致。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan/s/imlg0beiu)、[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5) |
+
+### 官方已下线或公告数日内下线（26 条）
+
+| 配置 | 型号 | 原因 |
+| --- | --- | --- |
+| `model-specs/moonshot.json` | `moonshot-v1-8k` | 官网模型列表：moonshot-v1 系列（含 -vision-preview）已于 2026-08-31 下线。来源：[platform.kimi.com](https://platform.kimi.com/docs/models.md) |
+| `model-specs/moonshot.json` | `moonshot-v1-32k` | 官网模型列表：moonshot-v1 系列（含 -vision-preview）已于 2026-08-31 下线。来源：[platform.kimi.com](https://platform.kimi.com/docs/models.md) |
+| `model-specs/moonshot.json` | `moonshot-v1-128k` | 官网模型列表：moonshot-v1 系列（含 -vision-preview）已于 2026-08-31 下线。来源：[platform.kimi.com](https://platform.kimi.com/docs/models.md) |
+| `model-specs/tencent.json` | `hunyuan-2.0-thinking-20251109` | 官方下线公告：2026-06-22 起正式下线；原腾讯混元大模型平台 2026-09-30 全面停服。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1729/131925) |
+| `model-specs/tencent.json` | `hunyuan-t1-latest` | 官方下线公告：2026-06-22 起正式下线；原腾讯混元大模型平台 2026-09-30 全面停服。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1729/131925) |
+| `providers/moonshot.json` | `moonshot-v1-8k` | 官网模型列表：moonshot-v1 系列（含 -vision-preview）已于 2026-08-31 下线。来源：[platform.kimi.com](https://platform.kimi.com/docs/models.md) |
+| `providers/moonshot.json` | `moonshot-v1-32k` | 官网模型列表：moonshot-v1 系列（含 -vision-preview）已于 2026-08-31 下线。来源：[platform.kimi.com](https://platform.kimi.com/docs/models.md) |
+| `providers/moonshot.json` | `moonshot-v1-128k` | 官网模型列表：moonshot-v1 系列（含 -vision-preview）已于 2026-08-31 下线。来源：[platform.kimi.com](https://platform.kimi.com/docs/models.md) |
+| `providers/moonshot.json` | `moonshot-v1-8k-vision-preview` | 官网模型列表：moonshot-v1 系列（含 -vision-preview）已于 2026-08-31 下线。来源：[platform.kimi.com](https://platform.kimi.com/docs/models.md) |
+| `providers/moonshot.json` | `moonshot-v1-32k-vision-preview` | 官网模型列表：moonshot-v1 系列（含 -vision-preview）已于 2026-08-31 下线。来源：[platform.kimi.com](https://platform.kimi.com/docs/models.md) |
+| `providers/moonshot.json` | `moonshot-v1-128k-vision-preview` | 官网模型列表：moonshot-v1 系列（含 -vision-preview）已于 2026-08-31 下线。来源：[platform.kimi.com](https://platform.kimi.com/docs/models.md) |
+| `providers/tencent.json` | `hunyuan-2.0-thinking-20251109` | 官方下线公告：2026-06-22 起正式下线；原腾讯混元大模型平台 2026-09-30 全面停服。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1729/131925) |
+| `providers/tencent.json` | `hunyuan-t1-latest` | 官方下线公告：2026-06-22 起正式下线；原腾讯混元大模型平台 2026-09-30 全面停服。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1729/131925) |
+| `coding-plans/volcengine-coding.json` | `doubao-seed-2.0-code` | Coding Plan 模型下线公告的「已下线模型」：08.08 下线。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-model-deprecation?lang=zh) |
+| `coding-plans/volcengine-coding.json` | `doubao-seed-2.0-pro` | Coding Plan 模型下线公告的「已下线模型」：08.08 下线。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-model-deprecation?lang=zh) |
+| `coding-plans/volcengine-coding.json` | `doubao-seed-2.0-lite` | Coding Plan 模型下线公告：2026-10-09 下线。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-model-deprecation?lang=zh) |
+| `coding-plans/volcengine-coding.json` | `doubao-seed-code` | Coding Plan 模型下线公告的「已下线模型」：08.05 下线。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-model-deprecation?lang=zh) |
+| `coding-plans/volcengine-coding.json` | `glm-4.7` | Coding Plan 模型下线公告的「已下线模型」：06.08 下线。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-model-deprecation?lang=zh) |
+| `coding-plans/volcengine-coding.json` | `deepseek-v3.2` | Coding Plan 模型下线公告的「已下线模型」：06.30 下线。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-model-deprecation?lang=zh) |
+| `coding-plans/volcengine-coding.json` | `kimi-k2.5` | Coding Plan 模型下线公告的「已下线模型」：06.08 下线。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-model-deprecation?lang=zh) |
+| `coding-plans/volcengine-coding.json` | `glm-5.1` | Coding Plan 模型下线公告的「已下线模型」：06.30 下线。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-model-deprecation?lang=zh) |
+| `coding-plans/volcengine-coding.json` | `kimi-k2.6` | Coding Plan 模型下线公告的「已下线模型」：08.18 下线。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-model-deprecation?lang=zh) |
+| `coding-plans/volcengine-coding.json` | `minimax-m2.5` | Coding Plan 模型下线公告的「已下线模型」：06.08 下线。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-model-deprecation?lang=zh) |
+| `coding-plans/volcengine-coding.json` | `minimax-m2.7` | Coding Plan 模型下线公告的「已下线模型」：08.18 下线。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/coding-plan-personal-model-deprecation?lang=zh) |
+| `coding-plans/tencent-token.json` | `glm-5.1` | TokenHub 模型列表的能力支持含 Function Calling，但套餐文档注明 2026-10-09 下线（模型列表写 2026-10-08）。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130060)、[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130051) |
+| `coding-plans/tencent-token.json` | `glm-5` | TokenHub 模型列表的能力支持含 Function Calling，但套餐文档注明 2026-10-09 下线（模型列表写 2026-10-08）。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130060)、[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130051) |
+
+### 官方当前文档里没有该型号，或没有它的工具调用说明（34 条）
+
+| 配置 | 型号 | 原因 |
+| --- | --- | --- |
+| `model-specs/minimax.json` | `M2-her` | 官网 M2-her 指南的参数表只有 model、messages、temperature、top_p、max_completion_tokens、stream，没有 tools。来源：[platform.minimax.cn](https://platform.minimax.cn/docs/guides/text-chat.md) |
+| `model-specs/minimax.json` | `MiniMax-Text-01` | 当前对话接口文档的 model 可选值不含 MiniMax-Text-01，tools 参数也没有按型号说明。来源：[platform.minimax.cn](https://platform.minimax.cn/docs/api-reference/text-post.md) |
+| `model-specs/xiaomi.json` | `mimo-x-flash-preview` | 官网模型列表页没有这个型号。来源：[mimo.mi.com](https://mimo.mi.com/docs/en-US/quick-start/summary/model) |
+| `model-specs/xiaomi.json` | `mimo-x-pro-preview` | 官网模型列表页没有这个型号。来源：[mimo.mi.com](https://mimo.mi.com/docs/en-US/quick-start/summary/model) |
+| `model-specs/baidu.json` | `ernie-4.5-turbo-128k` | 千帆 Function calling 文档的「支持模型范围」（2026-07-09）没有 ERNIE 4.5 Turbo。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5) |
+| `model-specs/baidu.json` | `ernie-4.5-turbo-20260402` | 千帆 Function calling 文档的「支持模型范围」（2026-07-09）没有 ERNIE 4.5 Turbo。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5) |
+| `model-specs/cohere.json` | `north-small-translate-1-0` | 官网型号页只描述机器翻译用途，没有工具调用内容。来源：[docs.cohere.com](https://docs.cohere.com/docs/north-small-translate-1.0.md) |
+| `model-specs/tencent.json` | `hunyuan-t1-vision` | 不在文生文下线清单里，但没有找到写明该型号支持工具调用的页面（接口文档的 Tools 说明只提到 hunyuan-turbos、hunyuan-t1、hunyuan-functioncall）；原混元平台 2026-09-30 全面停服。来源：[cloud.tencent.com](https://cloud.tencent.com/document/api/1729/105701)、[cloud.tencent.com](https://cloud.tencent.com/document/product/1729/131925) |
+| `model-specs/volcengine.json` | `doubao-seed-1.6-flash` | 方舟模型列表（2026-09-28）及其「工具调用能力」表里已经没有这个型号。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/model-list?lang=zh) |
+| `model-specs/volcengine.json` | `doubao-seed-1.6-lite` | 方舟模型列表（2026-09-28）及其「工具调用能力」表里已经没有这个型号。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/model-list?lang=zh) |
+| `model-specs/volcengine.json` | `doubao-seed-1.6-vision` | 方舟模型列表（2026-09-28）及其「工具调用能力」表里已经没有这个型号。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/model-list?lang=zh) |
+| `model-specs/volcengine.json` | `deepseek-v3.2` | 方舟模型列表（2026-09-28）及其「工具调用能力」表里已经没有这个型号。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/model-list?lang=zh) |
+| `providers/volcengine.json` | `doubao-seed-1.6-flash` | 方舟模型列表（2026-09-28）及其「工具调用能力」表里已经没有这个型号。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/model-list?lang=zh) |
+| `providers/volcengine.json` | `doubao-seed-1.6-lite` | 方舟模型列表（2026-09-28）及其「工具调用能力」表里已经没有这个型号。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/model-list?lang=zh) |
+| `providers/volcengine.json` | `doubao-seed-1.6-vision` | 方舟模型列表（2026-09-28）及其「工具调用能力」表里已经没有这个型号。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/model-list?lang=zh) |
+| `providers/volcengine.json` | `deepseek-v3.2` | 方舟模型列表（2026-09-28）及其「工具调用能力」表里已经没有这个型号。来源：[docs.volcengine.com](https://docs.volcengine.com/docs/ark/model-list?lang=zh) |
+| `providers/minimax.json` | `M2-her` | 官网 M2-her 指南的参数表只有 model、messages、temperature、top_p、max_completion_tokens、stream，没有 tools。来源：[platform.minimax.cn](https://platform.minimax.cn/docs/guides/text-chat.md) |
+| `providers/xiaomi.json` | `mimo-x-flash-preview` | 官网模型列表页没有这个型号。来源：[mimo.mi.com](https://mimo.mi.com/docs/en-US/quick-start/summary/model) |
+| `providers/xiaomi.json` | `mimo-x-pro-preview` | 官网模型列表页没有这个型号。来源：[mimo.mi.com](https://mimo.mi.com/docs/en-US/quick-start/summary/model) |
+| `providers/baidu.json` | `ernie-4.5-turbo-128k` | 千帆 Function calling 文档的「支持模型范围」（2026-07-09）没有 ERNIE 4.5 Turbo。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5) |
+| `providers/baidu.json` | `ernie-4.5-turbo-20260402` | 千帆 Function calling 文档的「支持模型范围」（2026-07-09）没有 ERNIE 4.5 Turbo。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5) |
+| `providers/tencent.json` | `hunyuan-t1-vision` | 不在文生文下线清单里，但没有找到写明该型号支持工具调用的页面（接口文档的 Tools 说明只提到 hunyuan-turbos、hunyuan-t1、hunyuan-functioncall）；原混元平台 2026-09-30 全面停服。来源：[cloud.tencent.com](https://cloud.tencent.com/document/api/1729/105701)、[cloud.tencent.com](https://cloud.tencent.com/document/product/1729/131925) |
+| `coding-plans/tencent-token.json` | `hunyuan-2.0-thinking` | 套餐文档当前的可用模型里没有这个型号。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130060) |
+| `coding-plans/tencent-token.json` | `hunyuan-t1` | 套餐文档当前的可用模型里没有这个型号。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130060) |
+| `coding-plans/tencent-token.json` | `hunyuan-turbos` | 套餐文档当前的可用模型里没有这个型号。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130060) |
+| `coding-plans/tencent-token.json` | `minimax-m2.5` | 套餐文档当前的可用模型里没有这个型号。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130060) |
+| `coding-plans/tencent-token.json` | `kimi-k2.5` | 套餐文档当前的可用模型里没有这个型号。来源：[cloud.tencent.com](https://cloud.tencent.com/document/product/1823/130060) |
+| `coding-plans/baidu-coding.json` | `deepseek-v4-flash` | 在 Coding Plan 可配置的 Model Name 里，但千帆 Function calling 文档的「支持模型范围」（2026-07-09）没有它。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan/s/imlg0beiu)、[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5) |
+| `coding-plans/baidu-coding.json` | `ernie-4.5-turbo-20260402` | 在 Coding Plan 可配置的 Model Name 里，但千帆 Function calling 文档的「支持模型范围」（2026-07-09）没有它。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan/s/imlg0beiu)、[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5) |
+| `coding-plans/baidu-coding.json` | `glm-5.1` | 在 Coding Plan 可配置的 Model Name 里，但千帆 Function calling 文档的「支持模型范围」（2026-07-09）没有它。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan/s/imlg0beiu)、[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan-docs/s/xm95lyys5) |
+| `coding-plans/baidu-coding.json` | `kimi-k2.6` | Coding Plan 文档可配置的 Model Name 里没有这个型号。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan/s/imlg0beiu) |
+| `coding-plans/baidu-coding.json` | `minimax-m2.7` | Coding Plan 文档可配置的 Model Name 里没有这个型号。来源：[cloud.baidu.com](https://cloud.baidu.com/doc/qianfan/s/imlg0beiu) |
+| `coding-plans/moorethread-coding.json` | `mt-coder` | 官方套餐指南只出现 GLM-4.7，没有 mt-coder 这个别名。来源：[docs.mthreads.com](https://docs.mthreads.com/kuaecloud/kuaecloud-doc-online/coding_plan/user_guide/) |
+| `coding-plans/kwai-coding.json` | `kwai-coder` | 官方接入指南使用的模型 ID 是 kat-coder-pro-v2.5，没有 kwai-coder。来源：[www.streamlake.ai](https://www.streamlake.ai/document/DOC/mg6k6nlp8j6qxicx4c9) |
+
+### 只在本仓未适配的接口上支持（2 条）
+
+| 配置 | 型号 | 原因 |
+| --- | --- | --- |
+| `model-specs/openai.json` | `gpt-4o-realtime` | 官网的精确 ID 是 gpt-4o-realtime-preview；function_calling 只在 Realtime 端点提供，Chat Completions / Responses 为 Not supported。客户端没有 Realtime 适配器，Provider 已 tombstone 该型号。来源：[developers.openai.com](https://developers.openai.com/api/docs/models/gpt-4o-realtime-preview.md) |
+| `providers/openai.json` | `gpt-realtime-2.1-mini` | 官网 Supported features 列有 function_calling，但只有 Realtime 端点为 Supported；本接入是 Chat Completions 协议，catalog 已标记该型号协议不受支持。来源：[developers.openai.com](https://developers.openai.com/api/docs/models/gpt-realtime-2.1-mini.md) |
+
+### 官网读取失败（9 条）
+
+| 配置 | 型号 | 原因 |
+| --- | --- | --- |
+| `model-specs/baichuan.json` | `Baichuan-M3-Plus` | 官网读取失败：curl、浏览器与抓取工具都在 TLS 握手阶段被对端断开，没有取得依据。来源：[platform.baichuan-ai.com](https://platform.baichuan-ai.com/docs/api) |
+| `model-specs/baichuan.json` | `Baichuan-M3` | 官网读取失败：curl、浏览器与抓取工具都在 TLS 握手阶段被对端断开，没有取得依据。来源：[platform.baichuan-ai.com](https://platform.baichuan-ai.com/docs/api) |
+| `model-specs/baichuan.json` | `Baichuan-M2-Plus` | 官网读取失败：curl、浏览器与抓取工具都在 TLS 握手阶段被对端断开，没有取得依据。来源：[platform.baichuan-ai.com](https://platform.baichuan-ai.com/docs/api) |
+| `model-specs/baichuan.json` | `Baichuan-M2` | 官网读取失败：curl、浏览器与抓取工具都在 TLS 握手阶段被对端断开，没有取得依据。来源：[platform.baichuan-ai.com](https://platform.baichuan-ai.com/docs/api) |
+| `providers/baichuan.json` | `Baichuan-M3-Plus` | 官网读取失败：curl、浏览器与抓取工具都在 TLS 握手阶段被对端断开，没有取得依据。来源：[platform.baichuan-ai.com](https://platform.baichuan-ai.com/docs/api) |
+| `providers/baichuan.json` | `Baichuan-M3` | 官网读取失败：curl、浏览器与抓取工具都在 TLS 握手阶段被对端断开，没有取得依据。来源：[platform.baichuan-ai.com](https://platform.baichuan-ai.com/docs/api) |
+| `providers/baichuan.json` | `Baichuan-M2-Plus` | 官网读取失败：curl、浏览器与抓取工具都在 TLS 握手阶段被对端断开，没有取得依据。来源：[platform.baichuan-ai.com](https://platform.baichuan-ai.com/docs/api) |
+| `providers/baichuan.json` | `Baichuan-M2` | 官网读取失败：curl、浏览器与抓取工具都在 TLS 握手阶段被对端断开，没有取得依据。来源：[platform.baichuan-ai.com](https://platform.baichuan-ai.com/docs/api) |
+| `coding-plans/infini-coding.json` | `deepseek-v3` | 官网本次读取失败（TLS 握手被断开）；2026-10-04 的重检记录其更新日志已删除该型号。来源：[docs.infini-ai.com](https://docs.infini-ai.com/changelog.html) |
+
+## 地域与接入面差异
+
+- `qwen-max`、`qwen-plus`、`qwen3-vl-plus`、`qwen3-vl-flash`：百炼型号页只在华北 2（北京）写支持 Function Calling，新加坡、法兰克福、弗吉尼亚、中国香港写不支持。共享规格按北京地域记；预置的 `provider-dashscope-001` 是北京兼容端点。用户自建的国际站接入会匹配到同一份规格，标签对它不适用。
+- `gpt-oss-120b`：OpenAI 直连只有 Responses 端点，Chat Completions 为 Not supported；其他托管平台另核。
+- `dashscope-coding` 的 `glm-4.7`：百炼 Function Calling 指南注明调用 GLM 系列要在请求里传 `tool_stream=true`，否则不返回 tool_calls。套餐端点是否同样要求没有核实。
+- `ark-code-latest`、`tc-code-latest`：别名本身没有能力表，依据是套餐文档加同平台模型列表，证据强度低于逐型号写明的条目。
+
+## 顺带看到、没有在本次处理的过期预置
+
+以下是核对过程中读到的官方下线信息，本次只据此决定不改标签，没有修改生命周期或移除条目：
+
+- 月之暗面：`moonshot-v1` 系列（含 vision-preview）与 `kimi-k2.5` 已于 2026-08-31 下线。
+- 腾讯混元：`hunyuan-2.0-thinking-20251109`、`hunyuan-t1-latest`、`hunyuan-turbos-latest` 等旧版文生文模型 2026-06-22 下线，原混元平台 2026-09-30 全面停服；Token Plan 当前可用模型里已经没有 `hunyuan-2.0-thinking`、`hunyuan-t1`、`hunyuan-turbos`、`minimax-m2.5`、`kimi-k2.5`，`glm-5` 与 `glm-5.1` 公告 2026-10-09 下线。
+- 火山方舟 Coding Plan：`doubao-seed-2.0-code`、`doubao-seed-2.0-pro`、`doubao-seed-code`、`glm-4.7`、`deepseek-v3.2`、`kimi-k2.5`、`glm-5.1`、`kimi-k2.6`、`minimax-m2.5`、`minimax-m2.7` 已下线，`doubao-seed-2.0-lite` 公告 2026-10-09 下线；方舟模型列表里也已经没有 `doubao-seed-1.6` 系列与 `deepseek-v3.2`。
+- 百度千帆：Coding Plan 自 2026-07-13 停止新购，由 Token Plan 个人版接替（接入地址不同）；模型列表把 `ernie-x1.1` 标为即将下线。
+- Perplexity：Sonar Chat Completions 的支持已于 2026-09-27 结束，同步与流式请求被改写为 Agent API 请求。