chore(internal-testing): 退役内置 provider；bump presetDataVersion to 61 (#26)

* chore(internal-testing): 退役内置 provider；bump presetDataVersion to 52

internal-testing 一直作为内测阶段的硬编码 provider 存在，现在客户端已切换到
登录后绑定 desirecore-cloud（NewAPI 网关）方案，internal-testing 不再需要。

从此 commit 起：
- compute/providers/internal-testing.json 删除
- _index.json 移除 internal-testing 条目
- presetDataVersion 51 → 52，触发客户端重新合并预置

客户端需配套实现 retired-providers 清理逻辑：在 mergeBuiltinIfNewer 之后
执行一次性清理，移除本地 compute.json 中残留的 provider-internal-testing-001
+ secrets.json#internal-testing。

* chore: bump preset data version to 61

* feat: 增量同步 16 个厂商旗舰模型 (首批精选)

首次从 OpenRouter 增量同步，覆盖 13 家厂商的最新旗舰：

概要
---
- 新增 .sync-watermark 水位线文件 (1782276303 / 2026-06-24)
- presetDataVersion: 61 → 62
- 校验 70/70 全部通过

新增模型
---
OpenAI:     gpt-image-2 (image_gen), gpt-5.5, gpt-5.4
Anthropic:  claude-fable-5, claude-opus-4.8
Google:     gemini-3.5-flash, gemini-3.1-flash-image (Nano Banana 2)
DeepSeek:   deepseek-v4-pro
Qwen:       qwen3.7-max, qwen3.7-plus
MiniMax:    MiniMax-M3
Moonshot:   kimi-k2.7-code
Zhipu:      glm-5.2
xAI:        grok-4.3
Mistral:    mistral-medium-3.5
Tencent:    hy3-preview
Cohere:     north-mini-code

后续机制
---
每日 10:00 心跳检查 OpenRouter 新模型，对比 .sync-watermark
增量入库，有更新才提 PR，无更新静默。

compute/model-specs/anthropic.json

@@ -1,6 +1,34 @@
 {
   "description": "Anthropic Claude 系列模型规格。参数来源：config-center compute/providers/anthropic.json。",
   "specs": [
+    {
+      "id": "claude-fable-5",
+      "displayName": "Claude Fable 5",
+      "family": "claude-fable",
+      "match": { "exact": ["claude-fable-5"], "patterns": ["claude-fable-5*"] },
+      "spec": {
+        "contextWindow": 1000000,
+        "maxOutputTokens": 128000,
+        "capabilities": ["chat", "reasoning", "deep_thinking", "code", "vision", "tool_use", "long_context"],
+        "serviceType": ["chat"],
+        "supportsReasoning": true,
+        "description": "Anthropic 最强模型，基于 Mythos 架构，支持自主知识工作"
+      }
+    },
+    {
+      "id": "claude-opus-4-8",
+      "displayName": "Claude Opus 4.8",
+      "family": "claude-opus",
+      "match": { "patterns": ["claude-opus-4-8*"] },
+      "spec": {
+        "contextWindow": 1000000,
+        "maxOutputTokens": 128000,
+        "capabilities": ["chat", "reasoning", "code", "vision", "tool_use", "long_context"],
+        "serviceType": ["chat"],
+        "supportsReasoning": true,
+        "description": "Claude Opus 4.8，Anthropic Opus 系列最新旗舰"
+      }
+    },
     {
       "id": "claude-opus-4-7",
       "displayName": "Claude Opus 4.7",

compute/model-specs/cohere.json

@@ -1,6 +1,20 @@
 {
   "description": "Cohere 系列模型规格。参数来源：config-center compute/providers/cohere.json。包含对话、Embedding 和 Rerank 三类服务。",
   "specs": [
+    {
+      "id": "north-mini-code",
+      "displayName": "North Mini Code",
+      "family": "north",
+      "match": { "exact": ["north-mini-code"], "patterns": ["north-mini-code*"] },
+      "spec": {
+        "contextWindow": 256000,
+        "maxOutputTokens": 64000,
+        "capabilities": ["chat", "code", "tool_use"],
+        "serviceType": ["chat"],
+        "supportsReasoning": true,
+        "description": "Cohere North 系列首发编程模型"
+      }
+    },
     {
       "id": "command-a-03-2025",
       "displayName": "Command A",

compute/model-specs/deepseek.json

@@ -1,6 +1,20 @@
 {
   "description": "DeepSeek 系列模型规格。参数来源：config-center compute/providers/deepseek.json。",
   "specs": [
+    {
+      "id": "deepseek-v4-pro",
+      "displayName": "DeepSeek V4 Pro",
+      "family": "deepseek-v4",
+      "match": { "exact": ["deepseek-v4-pro"], "patterns": ["deepseek-v4-pro*"] },
+      "spec": {
+        "contextWindow": 1048576,
+        "maxOutputTokens": 384000,
+        "capabilities": ["reasoning", "deep_thinking", "code", "math", "science", "multilingual", "tool_use"],
+        "serviceType": ["reasoning"],
+        "supportsReasoning": true,
+        "description": "DeepSeek V4 Pro，1.6T MoE 架构，100 万上下文"
+      }
+    },
     {
       "id": "deepseek-chat",
       "displayName": "DeepSeek Chat",

compute/model-specs/google.json

@@ -1,6 +1,34 @@
 {
   "description": "Google Gemini 系列模型规格。参数来源：config-center compute/providers/google.json。",
   "specs": [
+    {
+      "id": "gemini-3.5-flash",
+      "displayName": "Gemini 3.5 Flash",
+      "family": "gemini-3.5",
+      "match": { "exact": ["gemini-3.5-flash"], "patterns": ["gemini-3.5-flash*"] },
+      "spec": {
+        "contextWindow": 1048576,
+        "maxOutputTokens": 65536,
+        "capabilities": ["chat", "reasoning", "code", "vision", "video_understanding", "audio_understanding", "ultra_long_context", "tool_use", "fast"],
+        "serviceType": ["chat"],
+        "defaultTemperature": 1,
+        "supportsReasoning": true,
+        "description": "Google Gemini 3.5 Flash，全模态高性价比模型，100 万上下文"
+      }
+    },
+    {
+      "id": "gemini-3.1-flash-image",
+      "displayName": "Gemini 3.1 Flash Image (Nano Banana 2)",
+      "family": "gemini-3.1",
+      "match": { "exact": ["gemini-3.1-flash-image"], "patterns": ["gemini-3.1-flash-image*"] },
+      "spec": {
+        "contextWindow": 131072,
+        "maxOutputTokens": 32768,
+        "capabilities": ["image_generation", "image_editing", "text_rendering"],
+        "serviceType": ["image_gen"],
+        "description": "Google Nano Banana 2，最新文生图与编辑模型"
+      }
+    },
     {
       "id": "gemini-3.1-pro-preview",
       "displayName": "Gemini 3.1 Pro",

compute/model-specs/minimax.json

@@ -1,6 +1,21 @@
 {
   "description": "MiniMax 系列模型规格。参数来源：config-center compute/providers/minimax.json。id 保留上游 model_name 原始写法（含大小写与点号），匹配时由匹配器归一化处理。",
   "specs": [
+    {
+      "id": "MiniMax-M3",
+      "displayName": "MiniMax M3",
+      "family": "minimax-m3",
+      "match": { "exact": ["MiniMax-M3"], "patterns": ["MiniMax-M3*"] },
+      "spec": {
+        "contextWindow": 1048576,
+        "maxOutputTokens": 512000,
+        "capabilities": ["chat", "reasoning", "code", "vision", "video_understanding", "tool_use", "long_context"],
+        "serviceType": ["chat"],
+        "defaultTemperature": 1,
+        "supportsReasoning": true,
+        "description": "MiniMax M3 全模态基础模型，100 万上下文"
+      }
+    },
     {
       "id": "MiniMax-M2.7",
       "displayName": "MiniMax M2.7",

compute/model-specs/mistral.json

@@ -1,6 +1,20 @@
 {
   "description": "Mistral 系列模型规格。参数来源：config-center compute/providers/mistral.json。",
   "specs": [
+    {
+      "id": "mistral-medium-3.5",
+      "displayName": "Mistral Medium 3.5",
+      "family": "mistral-medium",
+      "match": { "exact": ["mistral-medium-3.5"], "patterns": ["mistral-medium-3.5*"] },
+      "spec": {
+        "contextWindow": 262144,
+        "capabilities": ["chat", "reasoning", "code", "vision", "tool_use"],
+        "serviceType": ["chat"],
+        "defaultTemperature": 1,
+        "supportsReasoning": true,
+        "description": "Mistral Medium 3.5，128B 指令模型"
+      }
+    },
     {
       "id": "mistral-large-latest",
       "displayName": "Mistral Large",

compute/model-specs/moonshot.json

@@ -1,6 +1,21 @@
 {
   "description": "Moonshot Kimi 系列模型规格。参数来源：config-center compute/providers/moonshot.json。注意 kimi-k2 存在多个子版本（k2.5/k2.6/k2-thinking），故 kimi-k2 仅用 exact 主键匹配，不用宽 pattern。",
   "specs": [
+    {
+      "id": "kimi-k2.7-code",
+      "displayName": "Kimi K2.7 Code",
+      "family": "kimi-k2.7",
+      "match": { "exact": ["kimi-k2.7-code"], "patterns": ["kimi-k2.7-code*"] },
+      "spec": {
+        "contextWindow": 262144,
+        "maxOutputTokens": 16384,
+        "capabilities": ["chat", "code", "vision", "tool_use"],
+        "serviceType": ["chat"],
+        "defaultTemperature": 1,
+        "supportsReasoning": true,
+        "description": "Moonshot Kimi K2.7 Code 编程专用模型"
+      }
+    },
     {
       "id": "kimi-k2.6",
       "displayName": "Kimi K2.6",

compute/model-specs/openai.json

@@ -1,6 +1,46 @@
 {
   "description": "OpenAI GPT / o 系列模型规格。参数来源：config-center compute/providers/openai.json。注意：gpt-4o / gpt-5 等存在子型号（-mini/-nano），故只用 exact 主键匹配，不使用宽泛 pattern 以防误吞子型号。o 系列 serviceType 由 responses 调整为 reasoning（responses 为 OpenAI 直连专有，对 NewAPI 网关不适用）。",
   "specs": [
+    {
+      "id": "gpt-image-2",
+      "displayName": "GPT Image 2",
+      "family": "gpt-image",
+      "match": { "exact": ["gpt-image-2"], "patterns": ["gpt-image-2*"] },
+      "spec": {
+        "contextWindow": 400000,
+        "capabilities": ["image_generation", "image_editing", "high_quality"],
+        "serviceType": ["image_gen"],
+        "description": "OpenAI 最新文生图模型，支持高质量图片生成与编辑，400K 上下文"
+      }
+    },
+    {
+      "id": "gpt-5.5",
+      "displayName": "GPT-5.5",
+      "family": "gpt-5",
+      "match": { "exact": ["gpt-5.5"], "patterns": ["gpt-5.5*"] },
+      "spec": {
+        "contextWindow": 1050000,
+        "maxOutputTokens": 128000,
+        "capabilities": ["chat", "reasoning", "code", "vision", "long_context", "tool_use"],
+        "serviceType": ["chat"],
+        "defaultTemperature": 1,
+        "supportsReasoning": true
+      }
+    },
+    {
+      "id": "gpt-5.4",
+      "displayName": "GPT-5.4",
+      "family": "gpt-5",
+      "match": { "exact": ["gpt-5.4"], "patterns": ["gpt-5.4*"] },
+      "spec": {
+        "contextWindow": 1050000,
+        "maxOutputTokens": 128000,
+        "capabilities": ["chat", "reasoning", "code", "vision", "long_context", "tool_use"],
+        "serviceType": ["chat"],
+        "defaultTemperature": 1,
+        "supportsReasoning": true
+      }
+    },
     {
       "id": "gpt-5.2",
       "displayName": "GPT-5.2",

compute/model-specs/qwen.json

@@ -1,6 +1,36 @@
 {
   "description": "阿里通义千问 Qwen 系列模型规格。参数来源：config-center compute/providers/dashscope.json。",
   "specs": [
+    {
+      "id": "qwen3.7-max",
+      "displayName": "Qwen3.7 Max",
+      "family": "qwen3.7",
+      "match": { "exact": ["qwen3.7-max"], "patterns": ["qwen3.7-max*"] },
+      "spec": {
+        "contextWindow": 1000000,
+        "maxOutputTokens": 65536,
+        "capabilities": ["chat", "reasoning", "code", "multilingual", "tool_use", "long_context"],
+        "serviceType": ["chat"],
+        "defaultTemperature": 1,
+        "supportsReasoning": true,
+        "description": "通义千问 Qwen3.7 Max 旗舰模型，100 万上下文"
+      }
+    },
+    {
+      "id": "qwen3.7-plus",
+      "displayName": "Qwen3.7 Plus",
+      "family": "qwen3.7",
+      "match": { "exact": ["qwen3.7-plus"], "patterns": ["qwen3.7-plus*"] },
+      "spec": {
+        "contextWindow": 1000000,
+        "maxOutputTokens": 65536,
+        "capabilities": ["chat", "reasoning", "code", "vision", "multilingual", "tool_use", "long_context"],
+        "serviceType": ["chat"],
+        "defaultTemperature": 1,
+        "supportsReasoning": true,
+        "description": "通义千问 Qwen3.7 Plus 高性价比多模态模型，100 万上下文"
+      }
+    },
     {
       "id": "qwen-max",
       "displayName": "Qwen Max",

compute/model-specs/tencent.json

@@ -1,6 +1,20 @@
 {
   "description": "腾讯混元系列模型规格。参数来源：config-center compute/providers/tencent.json。hunyuan-2.0-thinking 和 hunyuan-t1 为推理模型。",
   "specs": [
+    {
+      "id": "hy3-preview",
+      "displayName": "混元 Hy3 Preview",
+      "family": "hy3",
+      "match": { "exact": ["hy3-preview"], "patterns": ["hy3-preview*"] },
+      "spec": {
+        "contextWindow": 262144,
+        "capabilities": ["chat", "reasoning", "code", "multilingual", "tool_use"],
+        "serviceType": ["chat"],
+        "defaultTemperature": 1,
+        "supportsReasoning": true,
+        "description": "腾讯混元 Hy3 预览版，高效 MoE 模型"
+      }
+    },
     {
       "id": "hunyuan-2.0-thinking-20251109",
       "displayName": "混元 2.0 Think",

compute/model-specs/xai.json

@@ -1,6 +1,19 @@
 {
   "description": "xAI Grok 系列模型规格。参数来源：config-center compute/providers/xai.json。Grok 网关命名变体较多（含日期戳/-fast/-reasoning 后缀），主要依赖 family=grok-4 兜底匹配。",
   "specs": [
+    {
+      "id": "grok-4-3",
+      "displayName": "Grok 4.3",
+      "family": "grok-4",
+      "match": { "exact": ["grok-4.3"], "patterns": ["grok-4.3*"] },
+      "spec": {
+        "contextWindow": 1000000,
+        "capabilities": ["reasoning", "code", "vision", "tool_use", "long_context"],
+        "serviceType": ["chat"],
+        "supportsReasoning": true,
+        "description": "xAI Grok 4.3 推理模型，100 万上下文"
+      }
+    },
     {
       "id": "grok-4-1-fast-reasoning",
       "displayName": "Grok 4.1 Fast",

compute/model-specs/zhipu.json

@@ -1,6 +1,20 @@
 {
   "description": "智谱 GLM 系列模型规格。参数来源：config-center compute/providers/zhipu.json。glm-5 与 glm-5.1/glm-5-turbo/glm-5v-turbo 前缀相近，故各自仅用 exact 主键匹配，不用宽 pattern 以防误吞。",
   "specs": [
+    {
+      "id": "glm-5.2",
+      "displayName": "GLM-5.2",
+      "family": "glm-5.2",
+      "match": { "exact": ["glm-5.2"], "patterns": ["glm-5.2*"] },
+      "spec": {
+        "contextWindow": 1048576,
+        "maxOutputTokens": 32768,
+        "capabilities": ["reasoning", "deep_thinking", "code", "multilingual", "tool_use", "long_context"],
+        "serviceType": ["chat"],
+        "supportsReasoning": true,
+        "description": "智谱 GLM-5.2 大推理模型，100 万上下文"
+      }
+    },
     {
       "id": "glm-5.1",
       "displayName": "GLM-5.1",