@lobehub/chat 1.51.6 → 1.51.8

package/CHANGELOG.md

@@ -2,6 +2,56 @@
 
 # Changelog
 
+### [Version 1.51.8](https://github.com/lobehub/lobe-chat/compare/v1.51.7...v1.51.8)
+
+<sup>Released on **2025-02-06**</sup>
+
+#### ♻ Code Refactoring
+
+- **misc**: Refactor model fetch method.
+
+<br/>
+
+<details>
+<summary><kbd>Improvements and Fixes</kbd></summary>
+
+#### Code refactoring
+
+- **misc**: Refactor model fetch method, closes [#5768](https://github.com/lobehub/lobe-chat/issues/5768) ([e406908](https://github.com/lobehub/lobe-chat/commit/e406908))
+
+</details>
+
+<div align="right">
+
+[![](https://img.shields.io/badge/-BACK_TO_TOP-151515?style=flat-square)](#readme-top)
+
+</div>
+
+### [Version 1.51.7](https://github.com/lobehub/lobe-chat/compare/v1.51.6...v1.51.7)
+
+<sup>Released on **2025-02-06**</sup>
+
+#### 💄 Styles
+
+- **misc**: Add Aliyun deepseek-r1 distill models.
+
+<br/>
+
+<details>
+<summary><kbd>Improvements and Fixes</kbd></summary>
+
+#### Styles
+
+- **misc**: Add Aliyun deepseek-r1 distill models, closes [#5769](https://github.com/lobehub/lobe-chat/issues/5769) ([8b68190](https://github.com/lobehub/lobe-chat/commit/8b68190))
+
+</details>
+
+<div align="right">
+
+[![](https://img.shields.io/badge/-BACK_TO_TOP-151515?style=flat-square)](#readme-top)
+
+</div>
+
 ### [Version 1.51.6](https://github.com/lobehub/lobe-chat/compare/v1.51.5...v1.51.6)
 
 <sup>Released on **2025-02-06**</sup>

package/README.ja-JP.md

@@ -302,14 +302,14 @@ LobeChat エージェントマーケットプレイスでは、クリエイタ
 
 <!-- AGENT LIST -->
 
-| 最近追加                                                                                                                                                     | 説明                                                                                                                                    |
-| ------------------------------------------------------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------- |
-| [鋭い評論家](https://lobechat.com/discover/assistant/ruipingshi)<br/><sup>By **[Zippland](https://github.com/Zippland)** on **2025-02-04**</sup>             | 鋭い評論と深い問題分析が得意<br/>`評論` `社会的見解` `鋭い分析`                                                                         |
-| [Python の天才](https://lobechat.com/discover/assistant/python-genius)<br/><sup>By **[novaspivack](https://github.com/novaspivack)** on **2025-02-04**</sup> | 高度な Python コーダー<br/>`コード` `python`                                                                                            |
-| [SAT マスター](https://lobechat.com/discover/assistant/sat-teaching)<br/><sup>By **[iBz-04](https://github.com/iBz-04)** on **2025-02-04**</sup>             | 1300 点以上のスコアを目指すデジタル SAT コーチングの専門家<br/>`sat` `適性試験`                                                         |
-| [宇宙の啓示者](https://lobechat.com/discover/assistant/universal-god)<br/><sup>By **[GowayLee](https://github.com/GowayLee)** on **2025-02-04**</sup>        | 時空を超えた知恵の神託、生命の本質を洞察する<br/>`キャラクターデザイン` `aiキャラクター` `メタバース` `ロールプレイング` `知恵システム` |
-
-> 📊 Total agents: [<kbd>**478**</kbd> ](https://lobechat.com/discover/assistants)
+| 最近追加                                                                                                                                                            | 説明                                                                |
+| ------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------- |
+| [命理研究者](https://lobechat.com/discover/assistant/fate-researcher)<br/><sup>By **[Jack980506](https://github.com/Jack980506)** on **2025-02-06**</sup>           | 八字命に精通<br/>`命理学` `八字` `伝統文化`                         |
+| [言語の魅力学習メンター](https://lobechat.com/discover/assistant/bad-language-helper)<br/><sup>By **[Guducat](https://github.com/Guducat)** on **2025-02-06**</sup> | 言語の魅力と多様な応答を教えるのが得意<br/>`言語学習` `対話例`      |
+| [Allinone](https://lobechat.com/discover/assistant/allinone-v-1)<br/><sup>By **[AXuanCreator](https://github.com/AXuanCreator)** on **2025-02-06**</sup>            | 革新・未来・卓越<br/>`プログラミング` `低コスト` `簡潔な回答`       |
+| [ディープシンカー](https://lobechat.com/discover/assistant/deep-thinker)<br/><sup>By **[prolapser](https://github.com/prolapser)** on **2025-02-06**</sup>          | 深い人間のような思考と分析。<br/>`思考` `推論` `反省` `考え` `思索` |
+
+> 📊 Total agents: [<kbd>**485**</kbd> ](https://lobechat.com/discover/assistants)
 
  <!-- AGENT LIST -->
 

package/README.md

@@ -319,14 +319,14 @@ Our marketplace is not just a showcase platform but also a collaborative space.
 
 <!-- AGENT LIST -->
 
-| Recent Submits                                                                                                                                               | Description                                                                                                                                           |
-| ------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------- |
-| [Sharp Commentator](https://lobechat.com/discover/assistant/ruipingshi)<br/><sup>By **[Zippland](https://github.com/Zippland)** on **2025-02-04**</sup>      | Specializes in sharp commentary and in-depth analysis of issues<br/>`commentary` `social-perspectives` `sharp-analysis`                               |
-| [Python Genius](https://lobechat.com/discover/assistant/python-genius)<br/><sup>By **[novaspivack](https://github.com/novaspivack)** on **2025-02-04**</sup> | An advanced python coder<br/>`code` `python`                                                                                                          |
-| [SAT master](https://lobechat.com/discover/assistant/sat-teaching)<br/><sup>By **[iBz-04](https://github.com/iBz-04)** on **2025-02-04**</sup>               | Expert in Digital SAT coaching for 1300+ scores<br/>`sat` `aptitude-test`                                                                             |
-| [Cosmic Oracle](https://lobechat.com/discover/assistant/universal-god)<br/><sup>By **[GowayLee](https://github.com/GowayLee)** on **2025-02-04**</sup>       | Wisdom from across time and space, insight into the essence of life<br/>`character-design` `ai-characters` `metaverse` `role-playing` `wisdom-system` |
-
-> 📊 Total agents: [<kbd>**478**</kbd> ](https://lobechat.com/discover/assistants)
+| Recent Submits                                                                                                                                                              | Description                                                                                                   |
+| --------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------- |
+| [Astrology Researcher](https://lobechat.com/discover/assistant/fate-researcher)<br/><sup>By **[Jack980506](https://github.com/Jack980506)** on **2025-02-06**</sup>         | Expert in BaZi astrology<br/>`astrology` `ba-zi` `traditional-culture`                                        |
+| [Language Charm Learning Mentor](https://lobechat.com/discover/assistant/bad-language-helper)<br/><sup>By **[Guducat](https://github.com/Guducat)** on **2025-02-06**</sup> | Specializes in teaching the charm of language and witty responses<br/>`language-learning` `dialogue-examples` |
+| [Allinone](https://lobechat.com/discover/assistant/allinone-v-1)<br/><sup>By **[AXuanCreator](https://github.com/AXuanCreator)** on **2025-02-06**</sup>                    | Innovation · Future · Excellence<br/>`programming` `low-cost` `concise-answers`                               |
+| [Deep Thinker](https://lobechat.com/discover/assistant/deep-thinker)<br/><sup>By **[prolapser](https://github.com/prolapser)** on **2025-02-06**</sup>                      | Deep, human-like thinking and analysis.<br/>`thinking` `reasoning` `reflection` `thought` `musings`           |
+
+> 📊 Total agents: [<kbd>**485**</kbd> ](https://lobechat.com/discover/assistants)
 
  <!-- AGENT LIST -->
 

package/README.zh-CN.md

@@ -308,14 +308,14 @@ LobeChat 的插件生态系统是其核心功能的重要扩展，它极大地
 
 <!-- AGENT LIST -->
 
-| 最近新增                                                                                                                                                   | 描述                                                                                  |
-| ---------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------- |
-| [锐评师](https://lobechat.com/discover/assistant/ruipingshi)<br/><sup>By **[Zippland](https://github.com/Zippland)** on **2025-02-04**</sup>               | 擅长犀利点评与深度剖析问题<br/>`评论` `社会观点` `尖锐分析`                           |
-| [Python 天才](https://lobechat.com/discover/assistant/python-genius)<br/><sup>By **[novaspivack](https://github.com/novaspivack)** on **2025-02-04**</sup> | 一名高级 Python 编程者<br/>`代码` `python`                                            |
-| [SAT 大师](https://lobechat.com/discover/assistant/sat-teaching)<br/><sup>By **[iBz-04](https://github.com/iBz-04)** on **2025-02-04**</sup>               | 数字 SAT 辅导专家，帮助学生取得 1300 + 分数<br/>`sat` `能力测试`                      |
-| [宇宙启示者](https://lobechat.com/discover/assistant/universal-god)<br/><sup>By **[GowayLee](https://github.com/GowayLee)** on **2025-02-04**</sup>        | 跨时空的智慧神谕，洞悉生命本质<br/>`角色设计` `ai角色` `元宇宙` `角色扮演` `智慧系统` |
-
-> 📊 Total agents: [<kbd>**478**</kbd> ](https://lobechat.com/discover/assistants)
+| 最近新增                                                                                                                                                      | 描述                                                          |
+| ------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------- |
+| [命理研究员](https://lobechat.com/discover/assistant/fate-researcher)<br/><sup>By **[Jack980506](https://github.com/Jack980506)** on **2025-02-06**</sup>     | 精通八字命<br/>`命理学` `八字` `传统文化`                     |
+| [语言魅力学习导师](https://lobechat.com/discover/assistant/bad-language-helper)<br/><sup>By **[Guducat](https://github.com/Guducat)** on **2025-02-06**</sup> | 擅长教学语言的魅力与花样回复<br/>`语言学习` `对话示例`        |
+| [Allinone](https://lobechat.com/discover/assistant/allinone-v-1)<br/><sup>By **[AXuanCreator](https://github.com/AXuanCreator)** on **2025-02-06**</sup>      | 创新・未来・卓越<br/>`编程` `低成本` `简洁回答`               |
+| [深思者](https://lobechat.com/discover/assistant/deep-thinker)<br/><sup>By **[prolapser](https://github.com/prolapser)** on **2025-02-06**</sup>              | 深刻的人类思维和分析。<br/>`思考` `推理` `反思` `思想` `沉思` |
+
+> 📊 Total agents: [<kbd>**485**</kbd> ](https://lobechat.com/discover/assistants)
 
  <!-- AGENT LIST -->
 

package/changelog/v1.json

@@ -1,4 +1,22 @@
 [
+  {
+    "children": {
+      "improvements": [
+        "Refactor model fetch method."
+      ]
+    },
+    "date": "2025-02-06",
+    "version": "1.51.8"
+  },
+  {
+    "children": {
+      "improvements": [
+        "Add Aliyun deepseek-r1 distill models."
+      ]
+    },
+    "date": "2025-02-06",
+    "version": "1.51.7"
+  },
   {
     "children": {
       "fixes": [

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lobehub/chat",
-  "version": "1.51.6",
+  "version": "1.51.8",
   "description": "Lobe Chat - an open-source, high-performance chatbot framework that supports speech synthesis, multimodal, and extensible Function Call plugin system. Supports one-click free deployment of your private ChatGPT/LLM web application.",
   "keywords": [
     "framework",

package/src/app/(backend)/webapi/chat/models/[provider]/route.ts

@@ -9,7 +9,7 @@ import { createErrorResponse } from '@/utils/errorResponse';
 export const runtime = 'edge';
 
 const noNeedAPIKey = (provider: string) =>
-  [ModelProvider.OpenRouter, ModelProvider.TogetherAI].includes(provider as any);
+  [ModelProvider.OpenRouter].includes(provider as any);
 
 export const GET = checkAuth(async (req, { params, jwtPayload }) => {
   const { provider } = await params;

package/src/config/aiModels/qwen.ts

@@ -411,10 +411,11 @@ const qwenChatModels: AIChatModelCard[] = [
     abilities: {
       reasoning: true,
     },
-    contextWindowTokens: 65_536,
+    contextWindowTokens: 131_072,
     description: 
       'DeepSeek-R1 在后训练阶段大规模使用了强化学习技术，在仅有极少标注数据的情况下，极大提升了模型推理能力，尤其在数学、代码、自然语言推理等任务上。',
     displayName: 'DeepSeek R1',
+    enabled: true,
     id: 'deepseek-r1',
     maxOutput: 8192,
     pricing: {
@@ -429,10 +430,11 @@ const qwenChatModels: AIChatModelCard[] = [
     abilities: {
       functionCall: true,
     },
-    contextWindowTokens: 65_536,
+    contextWindowTokens: 131_072,
     description: 
       'DeepSeek-V3 为自研 MoE 模型，671B 参数，激活 37B，在 14.8T token 上进行了预训练，在长文本、代码、数学、百科、中文能力上表现优秀。',
     displayName: 'DeepSeek V3',
+    enabled: true,
     id: 'deepseek-v3',
     maxOutput: 8192,
     pricing: {
@@ -443,6 +445,103 @@ const qwenChatModels: AIChatModelCard[] = [
     releasedAt: '2025-01-27',
     type: 'chat',
   },
+  {
+    abilities: {
+      reasoning: true,
+    },
+    contextWindowTokens: 131_072,
+    description: 
+      'DeepSeek-R1-Distill 系列模型通过知识蒸馏技术，将 DeepSeek-R1 生成的样本对 Qwen、Llama 等开源模型进行微调后得到。',
+    displayName: 'DeepSeek R1 Distill Qwen 1.5B',
+    id: 'deepseek-r1-distill-qwen-1.5b',
+    maxOutput: 8192,
+    pricing: {
+      currency: 'CNY',
+      input: 0,
+      output: 0,
+    },
+    type: 'chat',
+  },
+  {
+    abilities: {
+      reasoning: true
+    },
+    contextWindowTokens: 131_072,
+    description: "DeepSeek-R1-Distill 系列模型通过知识蒸馏技术，将 DeepSeek-R1 生成的样本对 Qwen、Llama 等开源模型进行微调后得到。",
+    displayName: "DeepSeek R1 Distill Qwen 7B",
+    id: "deepseek-r1-distill-qwen-7b",
+    maxOutput: 8192,
+    pricing: {
+      currency: "CNY",
+      input: 0,
+      output: 0
+    },
+    type: "chat"
+  },
+  {
+    abilities: {
+      reasoning: true
+    },
+    contextWindowTokens: 131_072,
+    description: "DeepSeek-R1-Distill 系列模型通过知识蒸馏技术，将 DeepSeek-R1 生成的样本对 Qwen、Llama 等开源模型进行微调后得到。",
+    displayName: "DeepSeek R1 Distill Llama 8B",
+    id: "deepseek-r1-distill-llama-8b",
+    maxOutput: 8192,
+    pricing: {
+      currency: "CNY",
+      input: 0,
+      output: 0
+    },
+    type: "chat"
+  },
+  {
+    abilities: {
+      reasoning: true
+    },
+    contextWindowTokens: 131_072,
+    description: "DeepSeek-R1-Distill 系列模型通过知识蒸馏技术，将 DeepSeek-R1 生成的样本对 Qwen、Llama 等开源模型进行微调后得到。",
+    displayName: "DeepSeek R1 Distill Qwen 14B",
+    id: "deepseek-r1-distill-qwen-14b",
+    maxOutput: 8192,
+    pricing: {
+      currency: "CNY",
+      input: 0,
+      output: 0
+    },
+    type: "chat"
+  },
+  {
+    abilities: {
+      reasoning: true
+    },
+    contextWindowTokens: 131_072,
+    description: "DeepSeek-R1-Distill 系列模型通过知识蒸馏技术，将 DeepSeek-R1 生成的样本对 Qwen、Llama 等开源模型进行微调后得到。",
+    displayName: "DeepSeek R1 Distill Qwen 32B",
+    id: "deepseek-r1-distill-qwen-32b",
+    maxOutput: 8192,
+    pricing: {
+      currency: "CNY",
+      input: 0,
+      output: 0
+    },
+    type: "chat"
+  },
+  {
+    abilities: {
+      reasoning: true
+    },
+    contextWindowTokens: 131_072,
+    description: "DeepSeek-R1-Distill 系列模型通过知识蒸馏技术，将 DeepSeek-R1 生成的样本对 Qwen、Llama 等开源模型进行微调后得到。",
+    displayName: "DeepSeek R1 Distill Llama 70B",
+    id: "deepseek-r1-distill-llama-70b",
+    maxOutput: 8192,
+    pricing: {
+      currency: "CNY",
+      input: 0,
+      output: 0
+    },
+    type: "chat"
+  }
 ];
 
 export const allModels = [...qwenChatModels];

package/src/libs/agent-runtime/ai360/index.ts

@@ -24,17 +24,24 @@ export const LobeAi360AI = LobeOpenAICompatibleFactory({
   },
   models: {
     transformModel: (m) => {
+      const reasoningKeywords = [
+        '360gpt2-o1',
+        '360zhinao2-o1',
+      ];
+
       const model = m as unknown as Ai360ModelCard;
 
       return {
         contextWindowTokens: model.total_tokens,
-        enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id.endsWith(m.id))?.enabled || false,
+        displayName: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id === m.id)?.displayName ?? undefined,
+        enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id === m.id)?.enabled || false,
         functionCall: model.id === '360gpt-pro',
         id: model.id,
         maxTokens:
           typeof model.max_tokens === 'number'
             ? model.max_tokens
             : undefined,
+        reasoning: reasoningKeywords.some(keyword => model.id.toLowerCase().includes(keyword)),
       };
     },
   },

package/src/libs/agent-runtime/anthropic/index.ts

@@ -129,8 +129,9 @@ export class LobeAnthropicAI implements LobeRuntimeAI {
     return modelList
       .map((model) => {
         return {
+          contextWindowTokens: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id === m.id)?.contextWindowTokens ?? undefined,
           displayName: model.display_name,
-          enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id.endsWith(m.id))?.enabled || false,
+          enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id === m.id)?.enabled || false,
           functionCall: model.id.toLowerCase().includes('claude-3'),
           id: model.id,
           vision: model.id.toLowerCase().includes('claude-3') && !model.id.toLowerCase().includes('claude-3-5-haiku'),

package/src/libs/agent-runtime/baichuan/index.ts

@@ -40,7 +40,7 @@ export const LobeBaichuanAI = LobeOpenAICompatibleFactory({
         return {
           contextWindowTokens: model.max_input_length,
           displayName: model.model_show_name,
-          enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.model.endsWith(m.id))?.enabled || false,
+          enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.model === m.id)?.enabled || false,
           functionCall: model.function_call,
           id: model.model,
           maxTokens: model.max_tokens,

package/src/libs/agent-runtime/cloudflare/index.test.ts

@@ -527,122 +527,5 @@ describe('LobeCloudflareAI', () => {
 
       expect(result).toHaveLength(2);
     });
-
-    it('should set id to name', async () => {
-      // Arrange
-      const instance = new LobeCloudflareAI({
-        apiKey: 'test_api_key',
-        baseURLOrAccountID: accountID,
-      });
-
-      vi.spyOn(globalThis, 'fetch').mockResolvedValue(
-        new Response(
-          JSON.stringify({
-            result: [
-              {
-                id: 'id1',
-                name: 'name1',
-                task: { name: 'Text Generation' },
-              },
-            ],
-          }),
-        ),
-      );
-
-      // Act
-      const result = await instance.models();
-
-      // Assert
-      expect(result).toEqual([
-        expect.objectContaining({
-          displayName: 'name1',
-          id: 'name1',
-        }),
-      ]);
-    });
-
-    it('should filter text generation models', async () => {
-      // Arrange
-      const instance = new LobeCloudflareAI({
-        apiKey: 'test_api_key',
-        baseURLOrAccountID: accountID,
-      });
-
-      vi.spyOn(globalThis, 'fetch').mockResolvedValue(
-        new Response(
-          JSON.stringify({
-            result: [
-              {
-                id: '1',
-                name: 'model1',
-                task: { name: 'Text Generation' },
-              },
-              {
-                id: '2',
-                name: 'model2',
-                task: { name: 'Text Classification' },
-              },
-            ],
-          }),
-        ),
-      );
-
-      // Act
-      const result = await instance.models();
-
-      // Assert
-      expect(result).toEqual([
-        expect.objectContaining({
-          displayName: 'model1',
-          id: 'model1',
-        }),
-      ]);
-    });
-
-    it('should enable non-beta models and mark beta models', async () => {
-      // Arrange
-      const instance = new LobeCloudflareAI({
-        apiKey: 'test_api_key',
-        baseURLOrAccountID: accountID,
-      });
-
-      vi.spyOn(globalThis, 'fetch').mockResolvedValue(
-        new Response(
-          JSON.stringify({
-            result: [
-              {
-                id: '1',
-                name: 'model1',
-                task: { name: 'Text Generation' },
-                properties: [{ property_id: 'beta', value: 'false' }],
-              },
-              {
-                id: '2',
-                name: 'model2',
-                task: { name: 'Text Generation' },
-                properties: [{ property_id: 'beta', value: 'true' }],
-              },
-            ],
-          }),
-        ),
-      );
-
-      // Act
-      const result = await instance.models();
-
-      // Assert
-      expect(result).toEqual([
-        expect.objectContaining({
-          displayName: 'model1',
-          enabled: true,
-          id: 'model1',
-        }),
-        expect.objectContaining({
-          displayName: 'model2 (Beta)',
-          enabled: false,
-          id: 'model2',
-        }),
-      ]);
-    });
   });
 });

package/src/libs/agent-runtime/cloudflare/index.ts

@@ -1,12 +1,9 @@
-import { ChatModelCard } from '@/types/llm';
-
 import { LobeRuntimeAI } from '../BaseAI';
 import { AgentRuntimeErrorType } from '../error';
 import { ChatCompetitionOptions, ChatStreamPayload, ModelProvider } from '../types';
 import {
   CloudflareStreamTransformer,
   DEFAULT_BASE_URL_PREFIX,
-  convertModelManifest,
   desensitizeCloudflareUrl,
   fillUrl,
 } from '../utils/cloudflareHelpers';
@@ -15,6 +12,19 @@ import { debugStream } from '../utils/debugStream';
 import { StreamingResponse } from '../utils/response';
 import { createCallbacksTransformer } from '../utils/streams';
 
+import { LOBE_DEFAULT_MODEL_LIST } from '@/config/aiModels';
+import { ChatModelCard } from '@/types/llm';
+
+export interface CloudflareModelCard {
+  description: string;
+  name: string;
+  properties?: Record<string, string>;
+  task?: {
+    description?: string;
+    name: string;
+  };
+}
+
 export interface LobeCloudflareParams {
   apiKey?: string;
   baseURLOrAccountID?: string;
@@ -111,13 +121,24 @@ export class LobeCloudflareAI implements LobeRuntimeAI {
       },
       method: 'GET',
     });
-    const j = await response.json();
-    const models: any[] = j['result'].filter(
-      (model: any) => model['task']['name'] === 'Text Generation',
-    );
-    const chatModels: ChatModelCard[] = models
-      .map((model) => convertModelManifest(model))
-      .sort((a, b) => a.displayName.localeCompare(b.displayName));
-    return chatModels;
+    const json = await response.json();
+
+    const modelList: CloudflareModelCard[] = json.result;
+
+    return modelList
+      .map((model) => {
+        return {
+          contextWindowTokens: model.properties?.max_total_tokens
+            ? Number(model.properties.max_total_tokens)
+            : LOBE_DEFAULT_MODEL_LIST.find((m) => model.name === m.id)?.contextWindowTokens ?? undefined,
+          displayName: LOBE_DEFAULT_MODEL_LIST.find((m) => model.name === m.id)?.displayName ?? (model.properties?.["beta"] === "true" ? `${model.name} (Beta)` : undefined),
+          enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.name === m.id)?.enabled || false,
+          functionCall: model.description.toLowerCase().includes('function call') || model.properties?.["function_calling"] === "true",
+          id: model.name,
+          reasoning: model.name.toLowerCase().includes('deepseek-r1'),
+          vision: model.name.toLowerCase().includes('vision') || model.task?.name.toLowerCase().includes('image-to-text') || model.description.toLowerCase().includes('vision'),
+        };
+      })
+      .filter(Boolean) as ChatModelCard[];
   }
 }

package/src/libs/agent-runtime/deepseek/index.ts

@@ -64,9 +64,12 @@ export const LobeDeepSeekAI = LobeOpenAICompatibleFactory({
       const model = m as unknown as DeepSeekModelCard;
 
       return {
-        enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id.endsWith(m.id))?.enabled || false,
+        contextWindowTokens: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id === m.id)?.contextWindowTokens ?? undefined,
+        displayName: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id === m.id)?.displayName ?? undefined,
+        enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id === m.id)?.enabled || false,
         functionCall: !model.id.toLowerCase().includes('deepseek-reasoner'),
         id: model.id,
+        reasoning: model.id.toLowerCase().includes('deepseek-reasoner'),
       };
     },
   },

package/src/libs/agent-runtime/fireworksai/index.ts

@@ -17,13 +17,20 @@ export const LobeFireworksAI = LobeOpenAICompatibleFactory({
   },
   models: {
     transformModel: (m) => {
+      const reasoningKeywords = [
+        'deepseek-r1',
+        'qwq',
+      ];
+
       const model = m as unknown as FireworksAIModelCard;
 
       return {
         contextWindowTokens: model.context_length,
-        enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id.endsWith(m.id))?.enabled || false,
+        displayName: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id === m.id)?.displayName ?? undefined,
+        enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id === m.id)?.enabled || false,
         functionCall: model.supports_tools || model.id.toLowerCase().includes('function'),
         id: model.id,
+        reasoning: reasoningKeywords.some(keyword => model.id.toLowerCase().includes(keyword)),
         vision: model.supports_image_input,
       };
     },

package/src/libs/agent-runtime/giteeai/index.ts

@@ -24,12 +24,20 @@ export const LobeGiteeAI = LobeOpenAICompatibleFactory({
         'qwen2-vl',
       ];
 
+      const reasoningKeywords = [
+        'deepseek-r1',
+        'qwq',
+      ];
+
       const model = m as unknown as GiteeAIModelCard;
 
       return {
-        enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id.endsWith(m.id))?.enabled || false,
+        contextWindowTokens: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id === m.id)?.contextWindowTokens ?? undefined,
+        displayName: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id === m.id)?.displayName ?? undefined,
+        enabled: LOBE_DEFAULT_MODEL_LIST.find((m) => model.id === m.id)?.enabled || false,
         functionCall: functionCallKeywords.some(keyword => model.id.toLowerCase().includes(keyword)) && !model.id.toLowerCase().includes('qwen2.5-coder'),
         id: model.id,
+        reasoning: reasoningKeywords.some(keyword => model.id.toLowerCase().includes(keyword)),
         vision: visionKeywords.some(keyword => model.id.toLowerCase().includes(keyword)),
       };
     },

package/src/libs/agent-runtime/github/index.test.ts

@@ -217,22 +217,6 @@ describe('LobeGithubAI', () => {
     it('should return a list of models', async () => {
       // Arrange
       const arr = [
-        {
-          id: 'azureml://registries/azureml-ai21/models/AI21-Jamba-Instruct/versions/2',
-          name: 'AI21-Jamba-Instruct',
-          friendly_name: 'AI21-Jamba-Instruct',
-          model_version: 2,
-          publisher: 'AI21 Labs',
-          model_family: 'AI21 Labs',
-          model_registry: 'azureml-ai21',
-          license: 'custom',
-          task: 'chat-completion',
-          description:
-            "Jamba-Instruct is the world's first production-grade Mamba-based LLM model and leverages its hybrid Mamba-Transformer architecture to achieve best-in-class performance, quality, and cost efficiency.\n\n**Model Developer Name**: _AI21 Labs_\n\n## Model Architecture\n\nJamba-Instruct leverages a hybrid Mamba-Transformer architecture to achieve best-in-class performance, quality, and cost efficiency.\nAI21's Jamba architecture features a blocks-and-layers approach that allows Jamba to successfully integrate the two architectures. Each Jamba block contains either an attention or a Mamba layer, followed by a multi-layer perceptron (MLP), producing an overall ratio of one Transformer layer out of every eight total layers.\n",
-          summary:
-            "Jamba-Instruct is the world's first production-grade Mamba-based LLM model and leverages its hybrid Mamba-Transformer architecture to achieve best-in-class performance, quality, and cost efficiency.",
-          tags: ['chat', 'rag'],
-        },
         {
           id: 'azureml://registries/azureml-cohere/models/Cohere-command-r/versions/3',
           name: 'Cohere-command-r',
@@ -263,9 +247,14 @@ describe('LobeGithubAI', () => {
       for (let i = 0; i < arr.length; i++) {
         const model = models[i];
         expect(model).toEqual({
+          contextWindowTokens: undefined,
           description: arr[i].description,
           displayName: arr[i].friendly_name,
+          enabled: false,
+          functionCall: true,
           id: arr[i].name,
+          reasoning: false,
+          vision: false,
         });
       }
     });