codeep 2.16.0 → 2.18.0

package/dist/config/providers.js

@@ -18,14 +18,19 @@ export const PROVIDERS = {
             },
         },
         models: [
-            { id: 'glm-5.2', name: 'GLM-5.2', description: 'Latest GLM model, available to all users' },
+            // GLM-5.3 is Coding-Plan only — the standalone pay-per-use API does not
+            // accept it yet, so it must NOT be copied into `z.ai-api`.
+            { id: 'glm-5.3', name: 'GLM-5.3', description: 'Latest flagship for project-scale engineering (1M context)' },
+            { id: 'glm-5.2', name: 'GLM-5.2', description: 'Previous flagship for project-scale engineering (1M context)' },
             { id: 'glm-5-turbo', name: 'GLM-5 Turbo', description: 'Fast GLM-5 variant, available to all users' },
         ],
-        defaultModel: 'glm-5.2',
+        defaultModel: 'glm-5.3',
         defaultProtocol: 'openai',
+        maxOutputTokens: 131_072,
         envKey: 'ZAI_API_KEY',
         subscribeUrl: 'https://z.ai/subscribe?ic=NXYNXZOV14',
         groupLabel: 'Z.AI — Subscription (GLM Coding Plan)',
+        flatFee: true,
         hint: 'Uses your Z.AI subscription — no per-token charges.',
         mcpEndpoints: {
             webSearch: 'https://api.z.ai/api/mcp/web_search_prime/mcp',
@@ -44,11 +49,12 @@ export const PROVIDERS = {
             },
         },
         models: [
-            { id: 'glm-5.2', name: 'GLM-5.2', description: 'Latest GLM model' },
+            { id: 'glm-5.2', name: 'GLM-5.2', description: 'Latest flagship for project-scale engineering (1M context)' },
             { id: 'glm-5-turbo', name: 'GLM-5 Turbo', description: 'Fast GLM-5 variant' },
         ],
         defaultModel: 'glm-5.2',
         defaultProtocol: 'openai',
+        maxOutputTokens: 131_072,
         envKey: 'ZAI_API_KEY',
         subscribeUrl: 'https://api.z.ai',
         groupLabel: 'Z.AI — API (pay-per-use)',
@@ -70,14 +76,16 @@ export const PROVIDERS = {
             },
         },
         models: [
-            { id: 'glm-5.2', name: 'GLM-5.2', description: 'Latest GLM model, available to all users' },
+            { id: 'glm-5.2', name: 'GLM-5.2', description: 'Latest flagship for project-scale engineering (1M context)' },
             { id: 'glm-5-turbo', name: 'GLM-5 Turbo', description: 'Fast GLM-5 variant, available to all users' },
         ],
         defaultModel: 'glm-5.2',
         defaultProtocol: 'openai',
+        maxOutputTokens: 131_072,
         envKey: 'ZAI_CN_API_KEY',
         subscribeUrl: 'https://open.bigmodel.cn/glm-coding',
         groupLabel: 'Z.AI China — Subscription (GLM Coding Plan)',
+        flatFee: true,
         hint: 'Uses your ZhipuAI China subscription.',
         mcpEndpoints: {
             webSearch: 'https://open.bigmodel.cn/api/mcp/web_search_prime/mcp',
@@ -96,11 +104,12 @@ export const PROVIDERS = {
             },
         },
         models: [
-            { id: 'glm-5.2', name: 'GLM-5.2', description: 'Latest GLM model' },
+            { id: 'glm-5.2', name: 'GLM-5.2', description: 'Latest flagship for project-scale engineering (1M context)' },
             { id: 'glm-5-turbo', name: 'GLM-5 Turbo', description: 'Fast GLM-5 variant' },
         ],
         defaultModel: 'glm-5.2',
         defaultProtocol: 'openai',
+        maxOutputTokens: 131_072,
         envKey: 'ZAI_CN_API_KEY',
         subscribeUrl: 'https://open.bigmodel.cn',
         groupLabel: 'Z.AI China — API (pay-per-use)',
@@ -129,6 +138,7 @@ export const PROVIDERS = {
         envKey: 'MINIMAX_API_KEY',
         subscribeUrl: 'https://platform.minimax.io/subscribe/coding-plan?code=2lWvoWUhrp&source=link',
         groupLabel: 'MiniMax — Subscription',
+        flatFee: true,
         hint: 'Uses your MiniMax subscription — no per-token charges.',
     },
     'minimax-api': {
@@ -174,6 +184,7 @@ export const PROVIDERS = {
         envKey: 'MINIMAX_CN_API_KEY',
         subscribeUrl: 'https://platform.minimaxi.com',
         groupLabel: 'MiniMax China — Subscription',
+        flatFee: true,
         hint: 'Uses your MiniMax China subscription.',
     },
     'deepseek': {
@@ -205,9 +216,9 @@ export const PROVIDERS = {
     },
     // ── Kimi (Moonshot AI) ────────────────────────────────────────────
     // Subscription (Kimi Code) mirrors the Z.AI GLM-Coding-Plan shape: a
-    // dedicated coding base URL + a separate key, model id ALWAYS
-    // `kimi-for-coding` (a backend alias). OpenAI-compatible is the
-    // battle-tested path so we don't expose the Anthropic surface here.
+    // dedicated coding base URL + a separate key. Model availability depends
+    // on the user's plan, so the picker labels the restricted K3/high-speed ids
+    // instead of pretending every subscription includes them.
     'kimi': {
         name: 'Kimi (Moonshot) — Coding Plan',
         description: 'Kimi Code subscription',
@@ -215,7 +226,10 @@ export const PROVIDERS = {
             openai: { baseUrl: 'https://api.kimi.com/coding/v1', authHeader: 'Bearer', supportsNativeTools: true },
         },
         models: [
-            { id: 'kimi-for-coding', name: 'Kimi Code', description: 'Subscription alias — auto-maps to the latest Kimi coding model (K2.7 Code)' },
+            { id: 'kimi-for-coding', name: 'Kimi Code', description: 'Available on every Kimi Code plan — maps to K2.7 Code' },
+            { id: 'k3', name: 'Kimi K3', description: '1M-context flagship — Moderato plan or higher' },
+            { id: 'k3-256k', name: 'Kimi K3 (256K)', description: 'K3 with a smaller context window — Moderato plan or higher' },
+            { id: 'kimi-for-coding-highspeed', name: 'Kimi Code (High-Speed)', description: 'Low-latency K2.7 Code — Allegretto plan or higher' },
         ],
         defaultModel: 'kimi-for-coding',
         defaultProtocol: 'openai',
@@ -223,6 +237,7 @@ export const PROVIDERS = {
         envKey: 'KIMI_CODE_API_KEY',
         subscribeUrl: 'https://www.kimi.com/code',
         groupLabel: 'Kimi — Subscription (Kimi Code)',
+        flatFee: true,
         hint: 'Uses your Kimi Code subscription — no per-token charges. Key from kimi.com/code/console.',
     },
     'kimi-api': {
@@ -232,21 +247,18 @@ export const PROVIDERS = {
             openai: { baseUrl: 'https://api.moonshot.ai/v1', authHeader: 'Bearer', supportsNativeTools: true },
         },
         models: [
-            { id: 'kimi-k3-code', name: 'Kimi K3 Code', description: 'Newest flagship agentic coding model (1M context, deep reasoning)' },
-            { id: 'kimi-k3-code-highspeed', name: 'Kimi K3 Code (High-Speed)', description: 'Throughput-tuned K3 Code for latency-sensitive loops' },
-            { id: 'kimi-k3-thinking', name: 'Kimi K3 Thinking', description: 'K3 with explicit reasoning traces (highest quality, slower)' },
-            { id: 'kimi-k2.7-code', name: 'Kimi K2.7 Code', description: 'Previous-gen flagship agentic coding model (256K context)' },
+            { id: 'kimi-k3', name: 'Kimi K3', description: 'Latest flagship for software engineering and deep reasoning (1M context)' },
+            { id: 'kimi-k2.7-code', name: 'Kimi K2.7 Code', description: 'Coding-specialized model (256K context)' },
             { id: 'kimi-k2.7-code-highspeed', name: 'Kimi K2.7 Code (High-Speed)', description: 'Throughput-tuned K2.7 Code for latency-sensitive loops' },
-            { id: 'kimi-k2.6', name: 'Kimi K2.6', description: 'Previous-gen multimodal reasoning model' },
-            { id: 'kimi-k2.5', name: 'Kimi K2.5', description: 'Older general-purpose model (cheaper)' },
+            { id: 'kimi-k2.6', name: 'Kimi K2.6', description: 'General-purpose multimodal reasoning model' },
         ],
-        defaultModel: 'kimi-k3-code',
+        defaultModel: 'kimi-k3',
         defaultProtocol: 'openai',
-        maxOutputTokens: 65_536,
+        maxOutputTokens: 131_072,
         envKey: 'MOONSHOT_API_KEY',
         subscribeUrl: 'https://platform.kimi.ai/console/api-keys',
         groupLabel: 'Kimi — API (pay-per-use)',
-        hint: 'Pay-per-use via Moonshot API key (platform.kimi.ai). K3 models support 1M context and explicit reasoning.',
+        hint: 'Pay-per-use via Moonshot API key (platform.kimi.ai). Kimi K3 supports 1M context and graded reasoning.',
     },
     'kimi-cn': {
         name: 'Kimi China (Moonshot)',
@@ -255,21 +267,18 @@ export const PROVIDERS = {
             openai: { baseUrl: 'https://api.moonshot.cn/v1', authHeader: 'Bearer', supportsNativeTools: true },
         },
         models: [
-            { id: 'kimi-k3-code', name: 'Kimi K3 Code', description: 'Newest flagship agentic coding model (1M context, deep reasoning)' },
-            { id: 'kimi-k3-code-highspeed', name: 'Kimi K3 Code (High-Speed)', description: 'Throughput-tuned K3 Code' },
-            { id: 'kimi-k3-thinking', name: 'Kimi K3 Thinking', description: 'K3 with explicit reasoning traces' },
-            { id: 'kimi-k2.7-code', name: 'Kimi K2.7 Code', description: 'Previous-gen flagship agentic coding model (256K context)' },
+            { id: 'kimi-k3', name: 'Kimi K3', description: 'Latest flagship for software engineering and deep reasoning (1M context)' },
+            { id: 'kimi-k2.7-code', name: 'Kimi K2.7 Code', description: 'Coding-specialized model (256K context)' },
             { id: 'kimi-k2.7-code-highspeed', name: 'Kimi K2.7 Code (High-Speed)', description: 'Throughput-tuned K2.7 Code' },
-            { id: 'kimi-k2.6', name: 'Kimi K2.6', description: 'Previous-gen multimodal reasoning model' },
-            { id: 'kimi-k2.5', name: 'Kimi K2.5', description: 'Older general-purpose model' },
+            { id: 'kimi-k2.6', name: 'Kimi K2.6', description: 'General-purpose multimodal reasoning model' },
         ],
-        defaultModel: 'kimi-k3-code',
+        defaultModel: 'kimi-k3',
         defaultProtocol: 'openai',
-        maxOutputTokens: 65_536,
+        maxOutputTokens: 131_072,
         envKey: 'MOONSHOT_CN_API_KEY',
         subscribeUrl: 'https://platform.moonshot.cn/console/api-keys',
         groupLabel: 'Kimi China — API (pay-per-use)',
-        hint: 'Pay-per-use via Moonshot China API key (platform.moonshot.cn). K3 models support 1M context.',
+        hint: 'Pay-per-use via Moonshot China API key (platform.moonshot.cn). Kimi K3 supports 1M context.',
     },
     // ── Grok (xAI) ────────────────────────────────────────────────────
     // Pay-per-use today (console.x.ai key). The SuperGrok / X Premium+
@@ -282,12 +291,15 @@ export const PROVIDERS = {
             openai: { baseUrl: 'https://api.x.ai/v1', authHeader: 'Bearer', supportsNativeTools: true },
         },
         models: [
-            { id: 'grok-4.5', name: 'Grok 4.5', description: 'Flagship reasoning model — highest quality, 500K context' },
+            { id: 'grok-4.6', name: 'Grok 4.6', description: 'Flagship reasoning model — xAI recommends it for code, 500K context' },
+            { id: 'grok-4.5', name: 'Grok 4.5', description: 'Previous flagship reasoning model, 500K context' },
             { id: 'grok-build-0.1', name: 'Grok Build 0.1', description: 'Agentic coding model — fast, 256K context' },
-            { id: 'grok-4.3', name: 'Grok 4.3', description: 'Previous flagship, 1M context' },
-            { id: 'grok-code-fast-1', name: 'Grok Code Fast 1', description: 'Low-cost speed-first coder (alias of Build 0.1)' },
-            { id: 'grok-4-fast-reasoning', name: 'Grok 4 Fast (reasoning)', description: 'Cheap reasoning model, very large context' },
+            { id: 'grok-4.3', name: 'Grok 4.3', description: 'Older flagship, 1M context' },
         ],
+        // Stays on the agentic coder, not the new flagship. grok-4.6 is the better
+        // model and xAI recommends it for code, but it bills 2x input and 3x output
+        // against grok-build-0.1 — moving every unpinned user onto it silently is
+        // not ours to decide. It is one `/model` away for anyone who wants it.
         defaultModel: 'grok-build-0.1',
         defaultProtocol: 'openai',
         useMaxCompletionTokens: true, // reasoning models reject max_tokens
@@ -307,17 +319,18 @@ export const PROVIDERS = {
             openai: { baseUrl: 'https://coding-intl.dashscope.aliyuncs.com/v1', authHeader: 'Bearer', supportsNativeTools: true },
         },
         models: [
-            { id: 'qwen3-coder-plus', name: 'Qwen3-Coder Plus', description: 'Flagship coding model — best quality' },
-            { id: 'qwen3-coder-next', name: 'Qwen3-Coder Next', description: 'Balanced quality/speed/cost' },
-            { id: 'qwen3.7-max', name: 'Qwen3.7-Max', description: 'Flagship general model (code + reasoning)' },
+            { id: 'qwen3.7-plus', name: 'Qwen3.7-Plus', description: 'Recommended current model for coding, reasoning, and vision' },
+            { id: 'qwen3.6-plus', name: 'Qwen3.6-Plus', description: 'Fast multimodal model with a 1M context window' },
+            { id: 'qwen3.5-plus', name: 'Qwen3.5-Plus', description: 'Efficient general-purpose Coding Plan model' },
         ],
-        defaultModel: 'qwen3-coder-plus',
+        defaultModel: 'qwen3.7-plus',
         defaultProtocol: 'openai',
         maxOutputTokens: 65_536,
         noStreamWithTools: true,
         envKey: 'BAILIAN_CODING_PLAN_API_KEY',
         subscribeUrl: 'https://www.alibabacloud.com/help/en/model-studio/qwen-code-coding-plan',
         groupLabel: 'Qwen — Subscription (Coding Plan)',
+        flatFee: true,
         hint: 'Uses your Qwen Coding Plan — no per-token charges. sk-sp-… key from Model Studio. Interactive coding use only.',
     },
     'qwen-api': {
@@ -327,12 +340,11 @@ export const PROVIDERS = {
             openai: { baseUrl: 'https://dashscope-intl.aliyuncs.com/compatible-mode/v1', authHeader: 'Bearer', supportsNativeTools: true },
         },
         models: [
-            { id: 'qwen3-coder-plus', name: 'Qwen3-Coder Plus', description: 'Flagship coding model (256K, up to 1M)' },
-            { id: 'qwen3-coder-next', name: 'Qwen3-Coder Next', description: 'Balanced quality/speed/cost' },
-            { id: 'qwen3-coder-flash', name: 'Qwen3-Coder Flash', description: 'Fast/cheap coder' },
-            { id: 'qwen3.7-max', name: 'Qwen3.7-Max', description: 'Flagship general model' },
+            { id: 'qwen3.7-max', name: 'Qwen3.7-Max', description: 'Latest flagship for complex coding and reasoning' },
+            { id: 'qwen3.7-plus', name: 'Qwen3.7-Plus', description: 'Balanced quality, speed, and price (1M context)' },
+            { id: 'qwen3.6-flash', name: 'Qwen3.6-Flash', description: 'Low-latency, low-cost multimodal model' },
         ],
-        defaultModel: 'qwen3-coder-plus',
+        defaultModel: 'qwen3.7-max',
         defaultProtocol: 'openai',
         maxOutputTokens: 65_536,
         noStreamWithTools: true,
@@ -341,6 +353,29 @@ export const PROVIDERS = {
         groupLabel: 'Qwen — API (pay-per-use)',
         hint: 'Pay-per-use via Alibaba Model Studio key (DASHSCOPE_API_KEY).',
     },
+    'qwen-token-plan': {
+        name: 'Qwen (Alibaba) — Token Plan',
+        description: 'Qwen Token Plan subscription (international)',
+        protocols: {
+            openai: { baseUrl: 'https://token-plan.ap-southeast-1.maas.aliyuncs.com/compatible-mode/v1', authHeader: 'Bearer', supportsNativeTools: true },
+        },
+        models: [
+            { id: 'qwen3.8-max-preview', name: 'Qwen3.8-Max Preview', description: 'Newest Token Plan flagship for complex agentic work' },
+            { id: 'qwen3.7-max', name: 'Qwen3.7-Max', description: 'Production flagship for complex coding and reasoning' },
+            { id: 'qwen3.7-plus', name: 'Qwen3.7-Plus', description: 'Balanced quality and throughput' },
+            { id: 'qwen3.6-plus', name: 'Qwen3.6-Plus', description: 'Fast multimodal model with a 1M context window' },
+            { id: 'qwen3.6-flash', name: 'Qwen3.6-Flash', description: 'Low-latency, credit-efficient model' },
+        ],
+        defaultModel: 'qwen3.8-max-preview',
+        defaultProtocol: 'openai',
+        maxOutputTokens: 131_072,
+        noStreamWithTools: true,
+        envKey: 'BAILIAN_TOKEN_PLAN_API_KEY',
+        subscribeUrl: 'https://modelstudio.console.alibabacloud.com/',
+        groupLabel: 'Qwen — Subscription (Token Plan)',
+        flatFee: true,
+        hint: 'Uses monthly Token Plan credits. Requires a separate sk-sp-… Token Plan key.',
+    },
     'qwen-cn': {
         name: 'Qwen China — Coding Plan',
         description: 'Qwen Coding Plan subscription (China)',
@@ -348,17 +383,18 @@ export const PROVIDERS = {
             openai: { baseUrl: 'https://coding.dashscope.aliyuncs.com/v1', authHeader: 'Bearer', supportsNativeTools: true },
         },
         models: [
-            { id: 'qwen3-coder-plus', name: 'Qwen3-Coder Plus', description: 'Flagship coding model — best quality' },
-            { id: 'qwen3-coder-next', name: 'Qwen3-Coder Next', description: 'Balanced quality/speed/cost' },
-            { id: 'qwen3.7-max', name: 'Qwen3.7-Max', description: 'Flagship general model' },
+            { id: 'qwen3.7-plus', name: 'Qwen3.7-Plus', description: 'Recommended current model for coding, reasoning, and vision' },
+            { id: 'qwen3.6-plus', name: 'Qwen3.6-Plus', description: 'Fast multimodal model with a 1M context window' },
+            { id: 'qwen3.5-plus', name: 'Qwen3.5-Plus', description: 'Efficient general-purpose Coding Plan model' },
         ],
-        defaultModel: 'qwen3-coder-plus',
+        defaultModel: 'qwen3.7-plus',
         defaultProtocol: 'openai',
         maxOutputTokens: 65_536,
         noStreamWithTools: true,
         envKey: 'BAILIAN_CODING_PLAN_CN_API_KEY',
         subscribeUrl: 'https://bailian.console.aliyun.com/',
         groupLabel: 'Qwen China — Subscription (Coding Plan)',
+        flatFee: true,
         hint: 'Uses your Qwen Coding Plan (China). sk-sp-… key from Bailian.',
     },
     'qwen-cn-api': {
@@ -368,12 +404,11 @@ export const PROVIDERS = {
             openai: { baseUrl: 'https://dashscope.aliyuncs.com/compatible-mode/v1', authHeader: 'Bearer', supportsNativeTools: true },
         },
         models: [
-            { id: 'qwen3-coder-plus', name: 'Qwen3-Coder Plus', description: 'Flagship coding model' },
-            { id: 'qwen3-coder-next', name: 'Qwen3-Coder Next', description: 'Balanced quality/speed/cost' },
-            { id: 'qwen3-coder-flash', name: 'Qwen3-Coder Flash', description: 'Fast/cheap coder' },
-            { id: 'qwen3.7-max', name: 'Qwen3.7-Max', description: 'Flagship general model' },
+            { id: 'qwen3.7-max', name: 'Qwen3.7-Max', description: 'Latest flagship for complex coding and reasoning' },
+            { id: 'qwen3.7-plus', name: 'Qwen3.7-Plus', description: 'Balanced quality, speed, and price (1M context)' },
+            { id: 'qwen3.6-flash', name: 'Qwen3.6-Flash', description: 'Low-latency, low-cost multimodal model' },
         ],
-        defaultModel: 'qwen3-coder-plus',
+        defaultModel: 'qwen3.7-max',
         defaultProtocol: 'openai',
         maxOutputTokens: 65_536,
         noStreamWithTools: true,
@@ -384,21 +419,23 @@ export const PROVIDERS = {
     },
     'modelscope': {
         name: 'ModelScope (free Qwen)',
-        description: 'Free Qwen3-Coder inference via ModelScope',
+        description: 'Live free-model catalog via ModelScope',
         protocols: {
             openai: { baseUrl: 'https://api-inference.modelscope.cn/v1', authHeader: 'Bearer', supportsNativeTools: true },
         },
         models: [
-            { id: 'Qwen/Qwen3-Coder-480B-A35B-Instruct', name: 'Qwen3-Coder 480B', description: 'Open MoE coder — free tier (~2000 req/day)' },
+            { id: 'Qwen/Qwen3-Coder-480B-A35B-Instruct', name: 'Qwen3-Coder 480B', description: 'Fallback model shown until the live catalog loads' },
         ],
         defaultModel: 'Qwen/Qwen3-Coder-480B-A35B-Instruct',
         defaultProtocol: 'openai',
         maxOutputTokens: 65_536,
         noStreamWithTools: true,
+        dynamicModels: true,
         envKey: 'MODELSCOPE_API_KEY',
         subscribeUrl: 'https://modelscope.cn/my/myaccesstoken',
         groupLabel: 'ModelScope — Free (Qwen)',
-        hint: 'Free tier (~2000 req/day) via ModelScope token (modelscope.cn). Needs a bound Aliyun account.',
+        flatFee: true,
+        hint: 'Fetches the live free catalog for your ModelScope token; availability and limits vary by account.',
     },
     'openai': {
         name: 'OpenAI',
@@ -414,9 +451,6 @@ export const PROVIDERS = {
             { id: 'gpt-5.6-sol', name: 'GPT-5.6 Sol', description: 'Most capable GPT — best for coding & agentic work' },
             { id: 'gpt-5.6-terra', name: 'GPT-5.6 Terra', description: 'Balanced — GPT-5.5 quality at about half the price' },
             { id: 'gpt-5.6-luna', name: 'GPT-5.6 Luna', description: 'Fast and cheap — high-volume workloads' },
-            { id: 'gpt-5.5', name: 'GPT-5.5', description: 'Previous flagship GPT' },
-            { id: 'gpt-5.4', name: 'GPT-5.4', description: 'Older generation GPT' },
-            { id: 'gpt-5.4-mini', name: 'GPT-5.4 Mini', description: 'Faster and cheaper GPT-5.4' },
         ],
         defaultModel: 'gpt-5.6-sol',
         defaultProtocol: 'openai',
@@ -461,8 +495,10 @@ export const PROVIDERS = {
         },
         models: [
             { id: 'gemini-3.1-pro-preview', name: 'Gemini 3.1 Pro', description: 'Most capable Gemini model' },
-            { id: 'gemini-3.5-flash', name: 'Gemini 3.5 Flash', description: 'Pro-level intelligence at Flash speed' },
-            { id: 'gemini-3.1-flash-lite', name: 'Gemini 3.1 Flash-Lite', description: 'Low-latency, low-cost workhorse (1M context)' },
+            { id: 'gemini-3.7-flash', name: 'Gemini 3.7 Flash', description: 'Latest production Flash model — adjustable thinking, 64K output' },
+            { id: 'gemini-3.6-flash', name: 'Gemini 3.6 Flash', description: 'Previous production Flash model' },
+            { id: 'gemini-3.5-flash', name: 'Gemini 3.5 Flash', description: 'Stable frontier Flash model for coding and long agentic tasks' },
+            { id: 'gemini-3.5-flash-lite', name: 'Gemini 3.5 Flash-Lite', description: 'Latest low-latency, low-cost workhorse' },
         ],
         defaultModel: 'gemini-3.1-pro-preview',
         defaultProtocol: 'openai',
@@ -488,19 +524,17 @@ export const PROVIDERS = {
         models: [
             { id: 'openrouter/auto', name: 'Auto-route', description: 'OpenRouter picks the best model for the task' },
             { id: 'anthropic/claude-fable-5', name: 'Claude Fable 5', description: 'Anthropic — most capable' },
-            { id: 'anthropic/claude-opus-4', name: 'Claude Opus 4', description: 'Anthropic — Opus tier' },
+            { id: 'anthropic/claude-opus-5', name: 'Claude Opus 5', description: 'Anthropic — flagship Opus tier' },
             { id: 'anthropic/claude-sonnet-5', name: 'Claude Sonnet 5', description: 'Anthropic — balanced' },
             { id: 'openai/gpt-5.6-sol', name: 'GPT-5.6 Sol', description: 'OpenAI — flagship' },
-            { id: 'openai/gpt-5.4-mini', name: 'GPT-5.4 Mini', description: 'OpenAI — fast/cheap' },
-            { id: 'google/gemini-3.1-pro', name: 'Gemini 3.1 Pro', description: 'Google — multimodal' },
-            { id: 'meta-llama/llama-3.1-405b-instruct', name: 'Llama 3.1 405B', description: 'Meta — open weights, largest' },
-            { id: 'meta-llama/llama-3.1-70b-instruct', name: 'Llama 3.1 70B', description: 'Meta — open weights, fast' },
-            { id: 'deepseek/deepseek-v4', name: 'DeepSeek V4', description: 'DeepSeek via OpenRouter' },
-            { id: 'mistralai/mistral-large', name: 'Mistral Large', description: 'Mistral — flagship' },
-            { id: 'qwen/qwen-2.5-coder-32b-instruct', name: 'Qwen 2.5 Coder 32B', description: 'Alibaba — coding-tuned' },
-            { id: 'x-ai/grok-4.5', name: 'Grok 4.5', description: 'xAI — flagship reasoning' },
+            { id: 'openai/gpt-5.6-luna', name: 'GPT-5.6 Luna', description: 'OpenAI — fast/efficient' },
+            { id: 'google/gemini-3.7-flash', name: 'Gemini 3.7 Flash', description: 'Google — latest production Flash' },
+            { id: 'deepseek/deepseek-v4-pro', name: 'DeepSeek V4 Pro', description: 'DeepSeek — flagship agentic model' },
+            { id: 'moonshotai/kimi-k3', name: 'Kimi K3', description: 'Moonshot — long-horizon coding' },
+            { id: 'qwen/qwen3.8-max', name: 'Qwen 3.8 Max', description: 'Alibaba — latest flagship' },
+            { id: 'x-ai/grok-4.6', name: 'Grok 4.6', description: 'xAI — flagship reasoning' },
         ],
-        defaultModel: 'anthropic/claude-opus-4',
+        defaultModel: 'openrouter/auto',
         defaultProtocol: 'openai',
         envKey: 'OPENROUTER_API_KEY',
         subscribeUrl: 'https://openrouter.ai/keys',
@@ -550,6 +584,65 @@ export const PROVIDERS = {
 export function getProvider(id) {
     return PROVIDERS[id] || null;
 }
+/**
+ * Exact migrations for curated model ids that vendors replaced.
+ *
+ * Keep this deliberately narrower than the provider catalogue. Dynamic
+ * OpenRouter/Ollama/custom ids are user-controlled and must never be rewritten.
+ */
+const RETIRED_MODEL_REPLACEMENTS = {
+    'z.ai': { 'glm-5.1': 'glm-5.2', 'glm-5': 'glm-5.2' },
+    'z.ai-api': { 'glm-5.1': 'glm-5.2', 'glm-5': 'glm-5.2' },
+    'z.ai-cn': { 'glm-5.1': 'glm-5.2', 'glm-5': 'glm-5.2' },
+    'z.ai-cn-api': { 'glm-5.1': 'glm-5.2', 'glm-5': 'glm-5.2' },
+    google: {
+        'gemini-3.1-flash-lite': 'gemini-3.5-flash-lite',
+    },
+    grok: {
+        'grok-code-fast-1': 'grok-build-0.1',
+        'grok-4-fast-reasoning': 'grok-4.3',
+    },
+    openai: {
+        'gpt-5.5': 'gpt-5.6-sol',
+        'gpt-5.4': 'gpt-5.6-terra',
+        'gpt-5.4-mini': 'gpt-5.6-luna',
+    },
+    'kimi-api': {
+        'kimi-k3-code': 'kimi-k3',
+        'kimi-k3-code-highspeed': 'kimi-k3',
+        'kimi-k3-thinking': 'kimi-k3',
+        'kimi-k2.5': 'kimi-k2.6',
+    },
+    'kimi-cn': {
+        'kimi-k3-code': 'kimi-k3',
+        'kimi-k3-code-highspeed': 'kimi-k3',
+        'kimi-k3-thinking': 'kimi-k3',
+        'kimi-k2.5': 'kimi-k2.6',
+    },
+    qwen: {
+        'qwen3-coder-plus': 'qwen3.7-plus',
+        'qwen3-coder-next': 'qwen3.7-plus',
+        'qwen3.7-max': 'qwen3.7-plus',
+    },
+    'qwen-api': {
+        'qwen3-coder-plus': 'qwen3.7-max',
+        'qwen3-coder-next': 'qwen3.7-max',
+        'qwen3-coder-flash': 'qwen3.6-flash',
+    },
+    'qwen-cn': {
+        'qwen3-coder-plus': 'qwen3.7-plus',
+        'qwen3-coder-next': 'qwen3.7-plus',
+        'qwen3.7-max': 'qwen3.7-plus',
+    },
+    'qwen-cn-api': {
+        'qwen3-coder-plus': 'qwen3.7-max',
+        'qwen3-coder-next': 'qwen3.7-max',
+        'qwen3-coder-flash': 'qwen3.6-flash',
+    },
+};
+export function replacementModelFor(providerId, modelId) {
+    return RETIRED_MODEL_REPLACEMENTS[providerId]?.[modelId];
+}
 /**
  * Curated display order for the first-run login flow + `/provider` /
  * `/login` pickers. Headline / popular providers float to the top so
@@ -567,6 +660,7 @@ const DISPLAY_ORDER = [
     'kimi',
     'kimi-api',
     'qwen',
+    'qwen-token-plan',
     'qwen-api',
     'grok',
     'deepseek',
@@ -619,6 +713,13 @@ export function isNoApiKeyProvider(providerId) {
 export function isDynamicModelsProvider(providerId) {
     return PROVIDERS[providerId]?.dynamicModels === true;
 }
+/**
+ * Returns true if the provider bills a flat subscription (or is free), so any
+ * per-token dollar figure we compute for it is invented — see `flatFee`.
+ */
+export function isFlatFeeProvider(providerId) {
+    return PROVIDERS[providerId]?.flatFee === true;
+}
 export function getProviderBaseUrl(providerId, protocol) {
     const provider = PROVIDERS[providerId];
     if (!provider)
@@ -669,14 +770,23 @@ export function providerNoStreamWithTools(providerId) {
  * MODEL-level check, not a provider-level one (requiresDefaultTemperature
  * can't express it). Omitting the field is always safe — the API treats
  * omission as default. Kimi K2.x code/thinking models fix temperature
- * internally and 400 on any custom value, so they're here too.
+ * internally and 400 on any custom value, so they're here too. Google removed
+ * the deprecated sampling parameters outright in the Gemini 3.7 generation.
  */
 const SAMPLING_PARAMS_REJECTED = [
     'claude-fable-5', 'claude-opus-5', 'claude-opus-4-8', 'claude-opus-4-7', 'claude-sonnet-5',
-    'kimi-k3-code', 'kimi-k3-thinking', 'kimi-k2.7-code', 'kimi-for-coding',
+    'kimi-k3', 'kimi-k2.7-code', 'kimi-for-coding', 'k3',
+    'gemini-3.7-flash',
 ];
 export function modelRejectsSamplingParams(model) {
-    return SAMPLING_PARAMS_REJECTED.some(id => model === id || model.startsWith(`${id}-`));
+    // Canonicalize both sides. OpenRouter routes these as `google/gemini-3.7-flash`
+    // and `anthropic/claude-opus-4.8`, which a raw comparison misses — so the model
+    // gets a temperature it rejects, on the one path where the id is namespaced.
+    const id = canonicalModelId(model);
+    return SAMPLING_PARAMS_REJECTED.some(entry => {
+        const canonical = canonicalModelId(entry);
+        return id === canonical || id.startsWith(`${canonical}-`);
+    });
 }
 /**
  * Returns the effective max output tokens for a provider, capped by the provider's limit.
@@ -731,9 +841,14 @@ export function modelSupportsReasoningEffort(providerId, model) {
         case 'z.ai-api':
         case 'z.ai-cn':
         case 'z.ai-cn-api':
-            // GLM-5.2 added graded High/Max effort. glm-5-turbo is a plain thinking
-            // toggle (no graded levels) so it stays out.
-            return idMatches(id, 'glm-5-2');
+            // GLM-5.2 exposes graded High/Max effort; GLM-5.3 adds a distinct Low.
+            // Turbo is a plain toggle.
+            return idMatches(id, 'glm-5-2') || idMatches(id, 'glm-5-3');
+        case 'kimi':
+            return idMatches(id, 'k3');
+        case 'kimi-api':
+        case 'kimi-cn':
+            return idMatches(id, 'kimi-k3');
         case 'grok':
             // Grok reasoning models accept reasoning_effort (none/low/medium/high).
             // The coders (grok-code-fast, grok-build — the default) are NON-reasoning
@@ -743,8 +858,7 @@ export function modelSupportsReasoningEffort(providerId, model) {
             if (id.startsWith('grok-build') || id.startsWith('grok-code'))
                 return false;
             return id.startsWith('grok') && !id.includes('non-reasoning');
-        // Kimi (thinking on/off, not graded) and Qwen coders (non-thinking) have
-        // no graded knob → fall through to default false.
+        // GLM Turbo and Qwen coders expose thinking on/off, not a graded knob.
         case 'openrouter':
             // OpenRouter normalizes a unified `reasoning` field and silently ignores
             // it for non-reasoning models, so the control is always safe to expose.
@@ -775,15 +889,31 @@ export function reasoningParamsFor(providerId, model, tier) {
             // none/low/medium/high/xhigh — no "max"; map our Max → xhigh (the ceiling).
             return { reasoning_effort: tier === 'max' ? 'xhigh' : tier };
         case 'google':
-            // Gemini 3 (OpenAI-compat) accepts ONLY low/high — "medium" 400s.
-            return { reasoning_effort: tier === 'low' ? 'low' : 'high' };
+            // Gemini's OpenAI-compat layer maps reasoning_effort onto thinking_level
+            // and documents low | medium | high. (Medium 400'd on Gemini 3 Preview,
+            // which is why this used to collapse it — that was a preview-era bug and
+            // is fixed.) 'max' has no Gemini equivalent, so it tops out at high.
+            // 'minimal' is deliberately not emitted: 3.7 Flash rejects it outright.
+            return { reasoning_effort: tier === 'max' ? 'high' : tier };
         case 'deepseek':
+            // Graded thinking depth: high (default) or max. Lower tiers collapse to high.
+            return { reasoning_effort: tier === 'max' ? 'max' : 'high' };
         case 'z.ai':
         case 'z.ai-api':
         case 'z.ai-cn':
         case 'z.ai-cn-api':
-            // Graded thinking depth: high (default) or max. Lower tiers collapse to high.
+            // GLM-5.3 accepts low/high/max; GLM-5.2 grades only high|max, so lower
+            // tiers collapse to high there. Either way we always send an effort and
+            // never a disabled thinking block — GLM-5.3 rejects "disabled" outright.
+            if (idMatches(canonicalModelId(model), 'glm-5-3')) {
+                return { reasoning_effort: tier === 'low' ? 'low' : tier === 'max' ? 'max' : 'high' };
+            }
             return { reasoning_effort: tier === 'max' ? 'max' : 'high' };
+        case 'kimi':
+        case 'kimi-api':
+        case 'kimi-cn':
+            // Kimi K3 accepts low/high/max; collapse our medium tier to high.
+            return { reasoning_effort: tier === 'low' ? 'low' : tier === 'max' ? 'max' : 'high' };
         case 'grok':
             // none/low/medium/high — no "max"; map our Max → high (the ceiling).
             return { reasoning_effort: tier === 'max' ? 'high' : tier };
@@ -812,14 +942,24 @@ export function availableReasoningTiers(providerId, model) {
         case 'openai':
             return ['auto', 'low', 'medium', 'high', 'max'];
         case 'google':
-            // OpenAI-compat layer accepts only low/high — "medium" 400s.
-            return ['auto', 'low', 'high'];
+            // low | medium | high, per Gemini's OpenAI-compat mapping table. Medium
+            // is 3.7 Flash's own default and the tier Google recommends for agentic
+            // coding, so collapsing it hid the setting most users want.
+            return ['auto', 'low', 'medium', 'high'];
         case 'deepseek':
+            return ['auto', 'high', 'max'];
         case 'z.ai':
         case 'z.ai-api':
         case 'z.ai-cn':
         case 'z.ai-cn-api':
-            return ['auto', 'high', 'max'];
+            // GLM-5.3 distinguishes a Low tier; GLM-5.2 grades only high|max.
+            return idMatches(canonicalModelId(model), 'glm-5-3')
+                ? ['auto', 'low', 'high', 'max']
+                : ['auto', 'high', 'max'];
+        case 'kimi':
+        case 'kimi-api':
+        case 'kimi-cn':
+            return ['auto', 'low', 'high', 'max'];
         case 'grok':
             return ['auto', 'low', 'medium', 'high'];
         case 'openrouter':
@@ -831,8 +971,8 @@ export function availableReasoningTiers(providerId, model) {
 /**
  * Map a (possibly out-of-range) tier to the tier this model actually distinguishes,
  * for display — the chip + the checked menu row. The effort setting is global, so
- * a tier picked on Opus ('low') may not exist on GLM-5.2; we show the level GLM
- * will really run (its 'low' clamps to 'high'). Picks the available tier whose
+ * a tier picked on Opus ('medium') may not exist on Kimi K3; we show the level
+ * Kimi will really run (its 'medium' clamps to 'high'). Picks the tier whose
  * effective param equals the requested one. 'auto' (or unsupported) → 'auto'.
  */
 export function resolveReasoningTier(providerId, model, tier) {

package/dist/renderer/App.d.ts

@@ -89,6 +89,10 @@ export declare class App {
     private agentThinking;
     private agentWaitingForAI;
     private agentLog;
+    /** Process uptime shown in the persistent footer. */
+    private appStartedAt;
+    /** Start of the current agent run; unlike app uptime, resets per task. */
+    private agentStartedAt;
     private pasteInfo;
     private pasteInfoOpen;
     private codeBlockCounter;
@@ -433,6 +437,14 @@ export declare class App {
      * Render chat screen
      */
     private renderChat;
+    private shouldRenderAgentTimeline;
+    private renderPersistentHeader;
+    private renderAgentTimelineScreen;
+    private renderAgentContextRail;
+    private renderAgentKeyHints;
+    private timelineStatusStyle;
+    private currentActionType;
+    private currentAgentTask;
     /**
      * Render inline confirmation dialog below status bar
      */