desirecore/config-center
2
0
Fork 0
mirror of https://git.openapi.site/https://github.com/desirecore/config-center.git synced 2026-10-21 05:24:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix(compute): GLM-5.3 系列在 anthropic 协议按官网用 output_config.effort 控制思考强度 (#120)

Browse Source 
智谱官网（国内 coding-plan/latest-model、国际 devpack/latest-model）写明 /api/anthropic
以 thinking.type 与 output_config.effort 控制思考强度，thinking.type=enabled 一律按默认 max
处理，budget_tokens 不是控制项。两个 anthropic-messages 接入面的 GLM-5.3 系列此前只有按
OpenAI 协议文档写的 low/high/max 矩阵，客户端按 budget 形态发送：low/high 在上游都成了 max，
max（也是默认档）在客户端本机被拒绝；经 new-api 的 OpenAI 类渠道转换时 low 变成
reasoning_effort=medium，被智谱标准 API 拒绝。

- zhipu、zhipu-coding 接入面的 GLM-5.3 / 5.3-Flash / 5.3-FlashX 声明 adaptiveThinking
- GLM-5.3 系列规格加 anthropic-messages 协议 profile（V1 副本为 spec.extra.adaptiveThinking），
  覆盖经 ModelSpec 补全的云端条目
- 来源记录补国际 Coding Plan 页；按量 API Key 与 FlashX 按同一端点推定，未列入已核字段
- presetDataVersion 130
This commit is contained in:
yi-ge
2026-10-05 09:17:43 -07:00
committed by GitHub
parent e8122c1d77
commit 73b278df51
19 changed files with 170 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
compute/coding-plans/zhipu-coding.json
View File

@@ -50,6 +50,7 @@
"defaultTopP": 0.95,
"description": "智谱 GLM-5.3 大推理模型，面向复杂软件工程与长程 Agent 任务，100 万上下文；强制思考，无法关闭",
"extra": {
"adaptiveThinking": true,
"thinkingDefault": true,
"thinkingOnly": true,
"reasoning": {
@@ -92,6 +93,7 @@
"defaultTemperature": 1,
"defaultTopP": 0.95,
"extra": {
"adaptiveThinking": true,
"supportsThinking": true,
"thinkingDefault": true,
"thinkingOnly": true,

2
compute/model-specs/_index.json
View File

@@ -1,5 +1,5 @@
{
"description": "模型规格库加载顺序索引，以及智能路由三档的同域定义。模型级路由策略仅写入各 ModelSpec 的 routing 字段，不另建 smart-routing 目录。 数据批次 129：按官网补对话规格的 tool_use 标签，路由策略不变；此前批次 126 为官方字段复核与临时、历史规格边界。 精确身份治理：隔离Nano/实时/固定变体，identity-only不继承family参数；发布批次由manifest统一标记。",
"description": "模型规格库加载顺序索引，以及智能路由三档的同域定义。模型级路由策略仅写入各 ModelSpec 的 routing 字段，不另建 smart-routing 目录。 数据批次 130：GLM-5.3 系列在 anthropic-messages 协议上按智谱官网以 thinking.type 与 output_config.effort 控制思考强度（protocolProfiles 声明 adaptive 形态），路由策略不变；此前批次 129 按官网补对话规格的 tool_use 标签，批次 126 为官方字段复核与临时、历史规格边界。 精确身份治理：隔离Nano/实时/固定变体，identity-only不继承family参数；发布批次由manifest统一标记。",
"order": [
"anthropic",
"openai",

3
compute/model-specs/zhipu.json
View File

@@ -33,6 +33,7 @@
"supportsReasoning": true,
"description": "Z.ai GLM-5.3-FlashX，高速多模态推理模型，100 万上下文，始终开启推理",
"extra": {
"adaptiveThinking": true,
"thinkingOnly": true
}
},
@@ -84,6 +85,7 @@
"supportsReasoning": true,
"description": "智谱 GLM-5.3 大推理模型，面向复杂软件工程与长程 Agent 任务，100 万上下文；强制思考，无法关闭",
"extra": {
"adaptiveThinking": true,
"thinkingDefault": true,
"thinkingOnly": true
},
@@ -140,6 +142,7 @@
"supportsReasoning": true,
"description": "智谱 GLM-5.3-Flash 原生多模态模型，混合稀疏与线性注意力，面向高效编码与长程 Agent 任务，100 万上下文；强制思考，无法关闭",
"extra": {
"adaptiveThinking": true,
"thinkingDefault": true,
"thinkingOnly": true
},

3
compute/providers/zhipu.json
View File

@@ -48,6 +48,7 @@
],
"description": "Z.ai GLM-5.3-FlashX，高速多模态推理模型，100 万上下文，始终开启推理",
"extra": {
"adaptiveThinking": true,
"thinkingOnly": true,
"cacheHitPrice": 0.57,
"reasoning": {
@@ -89,6 +90,7 @@
"defaultTopP": 0.95,
"description": "智谱 GLM-5.3-Flash 原生多模态模型，混合稀疏与线性注意力，面向高效编码与长程 Agent 任务，100 万上下文；强制思考，无法关闭",
"extra": {
"adaptiveThinking": true,
"thinkingDefault": true,
"thinkingOnly": true,
"cacheHitPrice": 0.23,
@@ -128,6 +130,7 @@
"defaultTopP": 0.95,
"description": "智谱 GLM-5.3 大推理模型，面向复杂软件工程与长程 Agent 任务，100 万上下文；强制思考，无法关闭",
"extra": {
"adaptiveThinking": true,
"thinkingDefault": true,
"thinkingOnly": true,
"cacheHitPrice": 2,

43
compute/v2/catalog.json
View File

@@ -9342,7 +9342,17 @@
"metadata": {
"recordedExtra": {},
"sourceFile": "compute/model-specs/zhipu.json"
}
},
"protocolProfiles": [
{
"apiFormats": [
"anthropic-messages"
],
"request": {
"adaptiveThinking": true
}
}
]
},
{
"modelRef": "zhipu:glm-5.3",
@@ -9401,7 +9411,17 @@
"recordedFacts": {
"releasedAt": "2026-08-16"
}
}
},
"protocolProfiles": [
{
"apiFormats": [
"anthropic-messages"
],
"request": {
"adaptiveThinking": true
}
}
]
},
{
"modelRef": "zhipu:glm-5.3-flash",
@@ -9463,7 +9483,17 @@
"recordedFacts": {
"releasedAt": "2026-08-26"
}
}
},
"protocolProfiles": [
{
"apiFormats": [
"anthropic-messages"
],
"request": {
"adaptiveThinking": true
}
}
]
},
{
"modelRef": "zhipu:glm-5.2",
@@ -25139,6 +25169,7 @@
},
"overrides": {},
"request": {
"adaptiveThinking": true,
"constraints": {
"thinkingOnly": true
},
@@ -25204,6 +25235,7 @@
},
"overrides": {},
"request": {
"adaptiveThinking": true,
"constraints": {
"thinkingOnly": true
},
@@ -25271,6 +25303,7 @@
},
"overrides": {},
"request": {
"adaptiveThinking": true,
"constraints": {
"thinkingOnly": true
},
@@ -29991,6 +30024,7 @@
},
"overrides": {},
"request": {
"adaptiveThinking": true,
"constraints": {
"thinkingOnly": true
},
@@ -30056,6 +30090,7 @@
}
},
"request": {
"adaptiveThinking": true,
"constraints": {
"thinkingOnly": true
},
@@ -30304,5 +30339,5 @@
"authority": "catalog/v2/catalog.json",
"verification": "migration-preserves-existing-evidence"
},
"digest": "7ed3f47d0c41bec661509d5cb76d8130dbd4fb1e63d14eae0e73de5f00bcc4cb"
"digest": "8f0c6ee356bb44e60fb17d1dca04b012d47cae9cea9fd7890f874894ee74cdb5"
}

2
compute/v2/compatibility/coding-plans/zhipu-coding.json
View File

@@ -50,6 +50,7 @@
"defaultTopP": 0.95,
"description": "智谱 GLM-5.3 大推理模型，面向复杂软件工程与长程 Agent 任务，100 万上下文；强制思考，无法关闭",
"extra": {
"adaptiveThinking": true,
"thinkingDefault": true,
"thinkingOnly": true,
"reasoning": {
@@ -92,6 +93,7 @@
"defaultTemperature": 1,
"defaultTopP": 0.95,
"extra": {
"adaptiveThinking": true,
"supportsThinking": true,
"thinkingDefault": true,
"thinkingOnly": true,

3
compute/v2/compatibility/model-specs/zhipu.json
View File

@@ -33,6 +33,7 @@
"supportsReasoning": true,
"description": "Z.ai GLM-5.3-FlashX，高速多模态推理模型，100 万上下文，始终开启推理",
"extra": {
"adaptiveThinking": true,
"thinkingOnly": true
}
},
@@ -84,6 +85,7 @@
"supportsReasoning": true,
"description": "智谱 GLM-5.3 大推理模型，面向复杂软件工程与长程 Agent 任务，100 万上下文；强制思考，无法关闭",
"extra": {
"adaptiveThinking": true,
"thinkingDefault": true,
"thinkingOnly": true
},
@@ -140,6 +142,7 @@
"supportsReasoning": true,
"description": "智谱 GLM-5.3-Flash 原生多模态模型，混合稀疏与线性注意力，面向高效编码与长程 Agent 任务，100 万上下文；强制思考，无法关闭",
"extra": {
"adaptiveThinking": true,
"thinkingDefault": true,
"thinkingOnly": true
},

3
compute/v2/compatibility/providers/zhipu.json
View File

@@ -48,6 +48,7 @@
],
"description": "Z.ai GLM-5.3-FlashX，高速多模态推理模型，100 万上下文，始终开启推理",
"extra": {
"adaptiveThinking": true,
"thinkingOnly": true,
"cacheHitPrice": 0.57,
"reasoning": {
@@ -89,6 +90,7 @@
"defaultTopP": 0.95,
"description": "智谱 GLM-5.3-Flash 原生多模态模型，混合稀疏与线性注意力，面向高效编码与长程 Agent 任务，100 万上下文；强制思考，无法关闭",
"extra": {
"adaptiveThinking": true,
"thinkingDefault": true,
"thinkingOnly": true,
"cacheHitPrice": 0.23,
@@ -128,6 +130,7 @@
"defaultTopP": 0.95,
"description": "智谱 GLM-5.3 大推理模型，面向复杂软件工程与长程 Agent 任务，100 万上下文；强制思考，无法关闭",
"extra": {
"adaptiveThinking": true,
"thinkingDefault": true,
"thinkingOnly": true,
"cacheHitPrice": 2,

2
compute/v2/manifest.json
View File

@@ -3,5 +3,5 @@
"state": "candidate",
"requiredClientVersion": null,
"catalogPath": "compute/v2/catalog.json",
"digest": "7ed3f47d0c41bec661509d5cb76d8130dbd4fb1e63d14eae0e73de5f00bcc4cb"
"digest": "8f0c6ee356bb44e60fb17d1dca04b012d47cae9cea9fd7890f874894ee74cdb5"
}

10
compute/v2/resolver-parity.json
View File

@@ -2974,7 +2974,7 @@
{
"accessId": "provider-zhipu-001",
"apiModelId": "glm-5.3-flashx",
"digest": "ea21bf8d96f65888a4f6991ed37a7824894d94fd1baa8509ccd9a4c9cea5c51c",
"digest": "d4886f90bf9c291fe3c23183348e995ad9a09d469b235ef0dc662e133a62d350",
"selectable": true,
"effectiveApiFormat": "anthropic-messages",
"effectiveApiModelId": "glm-5.3-flashx",
@@ -2986,7 +2986,7 @@
{
"accessId": "provider-zhipu-001",
"apiModelId": "glm-5.3-flash",
"digest": "0e50b1f0393cb2fd782ffa161b65c6151e213dcd7f9f4eae8621e4fdb1f4915d",
"digest": "a06221f39d4d13993cf160510bfe6dc2be5c28d082f6df7bd2ff81ce403d5b63",
"selectable": true,
"effectiveApiFormat": "anthropic-messages",
"effectiveApiModelId": "glm-5.3-flash",
@@ -2998,7 +2998,7 @@
{
"accessId": "provider-zhipu-001",
"apiModelId": "glm-5.3",
"digest": "44f0f52f12351b78825f7e32930e8e9f81693a80a3a277c196c30f88498e500d",
"digest": "d9c73c7eac661e0f6a429ead1ea86fb8920875b71c9577ec01c2f9f253cecb7b",
"selectable": true,
"effectiveApiFormat": "anthropic-messages",
"effectiveApiModelId": "glm-5.3",
@@ -4100,7 +4100,7 @@
{
"accessId": "provider-zhipu-coding-001",
"apiModelId": "glm-5.3",
"digest": "215a2293b36c7ba03f525fa42914eb81425051375a8b2f629bc76affc3d04691",
"digest": "ac57a48657071e66a8d0253e284b5cc598ae9bcf85d63ecde8e9d3672d17e8f8",
"selectable": true,
"effectiveApiFormat": "anthropic-messages",
"effectiveApiModelId": "glm-5.3",
@@ -4112,7 +4112,7 @@
{
"accessId": "provider-zhipu-coding-001",
"apiModelId": "glm-5.3-flash",
"digest": "c254302f180a5afb2303be38a65eb61237b464a4e4456b56297dcc9c2d2da38f",
"digest": "b44eecd20c164d0e36e7a54700f9f9505deb5d435b54532c25af9d3d899fdfa3",
"selectable": true,
"effectiveApiFormat": "anthropic-messages",
"effectiveApiModelId": "glm-5.3-flash",