♻️ refactor(model): clear and add models (lobehub#3208)

package.json

@@ -113,9 +113,9 @@
     "@khmyznikov/pwa-install": "^0.3.9",
     "@lobehub/chat-plugin-sdk": "^1.32.4",
     "@lobehub/chat-plugins-gateway": "^1.9.0",
-    "@lobehub/icons": "^1.26.0",
+    "@lobehub/icons": "^1.27.0",
     "@lobehub/tts": "^1.24.3",
-    "@lobehub/ui": "^1.146.6",
+    "@lobehub/ui": "^1.146.9",
     "@microsoft/fetch-event-source": "^2.0.1",
     "@neondatabase/serverless": "^0.9.4",
     "@next/third-parties": "^14.2.4",

src/config/modelProviders/deepseek.ts

@@ -1,5 +1,6 @@
 import { ModelProviderCard } from '@/types/llm';
 
+// ref https://platform.deepseek.com/api-docs/pricing
 const DeepSeek: ModelProviderCard = {
   chatModels: [
     {

src/config/modelProviders/google.ts

@@ -43,42 +43,30 @@ const Google: ModelProviderCard = {
       vision: true,
     },
     {
-      description: 'The best model for scaling across a wide range of tasks. This is the latest model.',
+      description:
+        'The best model for scaling across a wide range of tasks. This is the latest model.',
       displayName: 'Gemini 1.0 Pro',
       id: 'gemini-1.0-pro-latest',
       maxOutput: 2048,
       tokens: 30_720 + 2048,
     },
     {
-      description: 'The best model for scaling across a wide range of tasks. This is a stable model that supports tuning.',
+      description:
+        'The best model for scaling across a wide range of tasks. This is a stable model that supports tuning.',
       displayName: 'Gemini 1.0 Pro 001 (Tuning)',
       functionCall: true,
       id: 'gemini-1.0-pro-001',
       maxOutput: 2048,
       tokens: 30_720 + 2048,
     },
     {
-      description: 'The best model for scaling across a wide range of tasks. Released April 9, 2024.',
+      description:
+        'The best model for scaling across a wide range of tasks. Released April 9, 2024.',
       displayName: 'Gemini 1.0 Pro 002 (Tuning)',
       id: 'gemini-1.0-pro-002',
       maxOutput: 2048,
       tokens: 30_720 + 2048,
     },
-    {
-      description: 'The most capable model for highly complex tasks',
-      displayName: 'Gemini 1.0 Ultra',
-      id: 'gemini-ultra',
-      maxOutput: 2048,
-      tokens: 32_768,
-    },
-    {
-      description: 'A legacy text-only model optimized for chat conversations',
-      displayName: 'PaLM 2 Chat (Legacy)',
-      id: 'chat-bison-001',
-      legacy: true,
-      maxOutput: 1024,
-      // tokens: 4096 + 1024, // none tokens test
-    }
   ],
   checkModel: 'gemini-1.5-flash',
   id: 'google',

src/config/modelProviders/groq.ts

@@ -4,7 +4,7 @@ import { ModelProviderCard } from '@/types/llm';
 const Groq: ModelProviderCard = {
   chatModels: [
     {
-      displayName: 'LLaMA3-3-70B',
+      displayName: 'LLaMA3 70B',
       enabled: true,
       functionCall: true,
       id: 'llama3-70b-8192',
@@ -18,14 +18,20 @@ const Groq: ModelProviderCard = {
       tokens: 32_768,
     },
     {
-      displayName: 'Gemma-7b-it',
-      enabled: true,
+      displayName: 'Gemma 7B',
       functionCall: true,
       id: 'gemma-7b-it',
       tokens: 8192,
     },
     {
-      displayName: 'LLaMA3-3-8B',
+      displayName: 'Gemma2 9B',
+      enabled: true,
+      functionCall: true,
+      id: 'gemma2-9b-it',
+      tokens: 8192,
+    },
+    {
+      displayName: 'LLaMA3 8B',
       enabled: true,
       functionCall: true,
       id: 'llama3-8b-8192',
@@ -37,7 +43,7 @@ const Groq: ModelProviderCard = {
       tokens: 4096,
     },
   ],
-  checkModel: 'gemma-7b-it',
+  checkModel: 'gemma2-9b-it',
   id: 'groq',
   name: 'Groq',
   proxyUrl: {

src/config/modelProviders/ollama.ts

@@ -1,5 +1,6 @@
 import { ModelProviderCard } from '@/types/llm';
 
+// ref https://ollama.com/library
 const Ollama: ModelProviderCard = {
   chatModels: [
     {
@@ -78,6 +79,28 @@ const Ollama: ModelProviderCard = {
       id: 'gemma:2b',
       tokens: 8192,
     },
+    {
+      displayName: 'Deepseek V2 16B',
+      enabled: true,
+      id: 'deepseek-v2',
+      tokens: 32_000,
+    },
+    {
+      displayName: 'Deepseek V2 236B',
+      id: 'deepseek-v2:236b',
+      tokens: 128_000,
+    },
+    {
+      displayName: 'Deepseek Coder V2 16B', // https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct
+      enabled: true,
+      id: 'deepseek-coder-v2',
+      tokens: 128_000,
+    },
+    {
+      displayName: 'Deepseek Coder V2 236B', // https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct
+      id: 'deepseek-coder-v2:236b',
+      tokens: 128_000,
+    },
     {
       displayName: 'Llama2 Chat 13B',
       id: 'llama2:13b',
@@ -135,10 +158,10 @@ const Ollama: ModelProviderCard = {
       tokens: 16_384,
     },
     {
-      displayName: 'Mistral',
+      displayName: 'MathΣtral',
       enabled: true,
-      id: 'mistral',
-      tokens: 32_768, // https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2/blob/main/config.json
+      id: 'mathstral',
+      tokens: 32_000, // https://huggingface.co/mistralai/mathstral-7B-v0.1
     },
     {
       displayName: 'Mixtral 8x7B',

src/config/modelProviders/openai.ts

@@ -1,6 +1,8 @@
 import { ModelProviderCard } from '@/types/llm';
 
-// ref https://platform.openai.com/docs/models
+// ref:
+// https://platform.openai.com/docs/models
+// https://platform.openai.com/docs/deprecations
 const OpenAI: ModelProviderCard = {
   chatModels: [
     {
@@ -53,14 +55,14 @@ const OpenAI: ModelProviderCard = {
       tokens: 128_000,
     },
     {
-      description: 'Currently points to gpt-4-1106-vision-preview',
+      description: 'Currently points to gpt-4-1106-vision-preview', // Will be discontinued on December 6, 2024
       displayName: 'GPT-4 Turbo Vision Preview',
       id: 'gpt-4-vision-preview',
       tokens: 128_000,
       vision: true,
     },
     {
-      displayName: 'GPT-4 Turbo Vision Preview (1106)',
+      displayName: 'GPT-4 Turbo Vision Preview (1106)', // Will be discontinued on December 6, 2024
       id: 'gpt-4-1106-vision-preview',
       tokens: 128_000,
       vision: true,
@@ -85,20 +87,21 @@ const OpenAI: ModelProviderCard = {
       tokens: 8192,
     },
     {
-      description: 'Currently points to gpt-4-32k-0613',
+      description: 'Currently points to gpt-4-32k-0613', // Will be discontinued on June 6, 2025
       displayName: 'GPT-4 32K',
       functionCall: true,
       id: 'gpt-4-32k',
       tokens: 32_768,
     },
     {
-      displayName: 'GPT-4 32K (0613)',
+      displayName: 'GPT-4 32K (0613)', // Will be discontinued on June 6, 2025
       functionCall: true,
       id: 'gpt-4-32k-0613',
       tokens: 32_768,
     },
     {
-      description: 'GPT 3.5 Turbo，适用于各种文本生成和理解任务',
+      description:
+        'GPT 3.5 Turbo，适用于各种文本生成和理解任务，Currently points to gpt-3.5-turbo-0125',
       displayName: 'GPT-3.5 Turbo',
       functionCall: true,
       id: 'gpt-3.5-turbo',
@@ -122,20 +125,20 @@ const OpenAI: ModelProviderCard = {
       tokens: 4096,
     },
     {
-      description: 'Currently points to gpt-3.5-turbo-16k-0613',
+      description: 'Currently points to gpt-3.5-turbo-16k-0613', // Will be discontinued on September 13, 2024
       displayName: 'GPT-3.5 Turbo 16K',
       id: 'gpt-3.5-turbo-16k',
       legacy: true,
       tokens: 16_385,
     },
     {
-      displayName: 'GPT-3.5 Turbo (0613)',
+      displayName: 'GPT-3.5 Turbo (0613)', // Will be discontinued on September 13, 2024
       id: 'gpt-3.5-turbo-0613',
       legacy: true,
       tokens: 4096,
     },
     {
-      displayName: 'GPT-3.5 Turbo 16K (0613)',
+      description: 'Currently points to gpt-3.5-turbo-16k-0613', // Will be discontinued on September 13, 2024
       id: 'gpt-3.5-turbo-16k-0613',
       legacy: true,
       tokens: 16_385,

src/config/modelProviders/perplexity.ts

@@ -4,7 +4,7 @@ import { ModelProviderCard } from '@/types/llm';
 const Perplexity: ModelProviderCard = {
   chatModels: [
     {
-      displayName: 'Perplexity 7B Chat',
+      displayName: 'Perplexity 8B Chat',
       id: 'llama-3-sonar-small-32k-chat',
       tokens: 32_768,
     },
@@ -15,7 +15,7 @@ const Perplexity: ModelProviderCard = {
       tokens: 32_768,
     },
     {
-      displayName: 'Perplexity 7B Online',
+      displayName: 'Perplexity 8B Online',
       id: 'llama-3-sonar-small-32k-online',
       tokens: 28_000,
     },
@@ -41,7 +41,7 @@ const Perplexity: ModelProviderCard = {
       tokens: 16_384,
     },
   ],
-  checkModel: 'pplx-7b-chat',
+  checkModel: 'llama-3-8b-instruct',
   id: 'perplexity',
   name: 'Perplexity',
   proxyUrl: {

src/config/modelProviders/togetherai.ts

@@ -58,9 +58,9 @@ const TogetherAI: ModelProviderCard = {
       tokens: 32_768,
     },
     {
-      displayName: 'Qwen 1.5 Chat (14B)',
+      displayName: 'Qwen 1.5 Chat (32B)',
       enabled: true,
-      id: 'Qwen/Qwen1.5-14B-Chat',
+      id: 'Qwen/Qwen1.5-32B-Chat',
       tokens: 32_768,
     },
     {

src/config/modelProviders/zeroone.ts

@@ -11,7 +11,8 @@ const ZeroOne: ModelProviderCard = {
       tokens: 32_768,
     },
     {
-      description: '在 yi-large 模型的基础上支持并强化了工具调用的能力，适用于各种需要搭建 agent 或 workflow 的业务场景。',
+      description:
+        '在 yi-large 模型的基础上支持并强化了工具调用的能力，适用于各种需要搭建 agent 或 workflow 的业务场景。',
       displayName: 'Yi Large FC',
       enabled: true,
       functionCall: true,
@@ -67,14 +68,9 @@ const ZeroOne: ModelProviderCard = {
       id: 'yi-large-preview',
       tokens: 16_384,
     },
-    {
-      description: '「兼容版本模型」实时信息获取，以及文本推理能力增强。',
-      displayName: 'Yi Large RAG Preview',
-      id: 'yi-large-rag-preview',
-      tokens: 16_384,
-    },
   ],
   checkModel: 'yi-large',
+  disableBrowserRequest: true,
   id: 'zeroone',
   name: '01.AI',
 };

src/config/modelProviders/zhipu.ts

@@ -6,7 +6,8 @@ import { ModelProviderCard } from '@/types/llm';
 const ZhiPu: ModelProviderCard = {
   chatModels: [
     {
-      description: 'GLM-4-AllTools 是专门为支持智能体和相关任务而进一步优化的模型版本。它能够自主理解用户的意图，规划复杂的指令，并能够调用一个或多个工具（例如网络浏览器、代码解释器和文本生图像）以完成复杂的任务。',
+      description:
+        'GLM-4-AllTools 是专门为支持智能体和相关任务而进一步优化的模型版本。它能够自主理解用户的意图，规划复杂的指令，并能够调用一个或多个工具（例如网络浏览器、代码解释器和文本生图像）以完成复杂的任务。',
       displayName: 'GLM-4-AllTools',
       enabled: true,
       functionCall: true,
@@ -50,22 +51,25 @@ const ZhiPu: ModelProviderCard = {
       tokens: 128_000,
     },
     {
-      description: '实现了视觉语言特征的深度融合，支持视觉问答、图像字幕、视觉定位、复杂目标检测等各类图像理解任务',
+      description:
+        '实现了视觉语言特征的深度融合，支持视觉问答、图像字幕、视觉定位、复杂目标检测等各类图像理解任务',
       displayName: 'GLM-4V',
       enabled: true,
       id: 'glm-4v',
       tokens: 2000,
       vision: true,
     },
     {
-      description: '适用于对知识量、推理能力、创造力要求较高的场景，比如广告文案、小说写作、知识类写作、代码生成等',
+      description:
+        '适用于对知识量、推理能力、创造力要求较高的场景，比如广告文案、小说写作、知识类写作、代码生成等', // Will be discontinued on December 31, 2024
       displayName: 'GLM-3-Turbo',
       functionCall: true,
       id: 'glm-3-turbo',
       tokens: 128_000,
     },
     {
-      description: 'CodeGeeX是一款强大的AI编程助手，提供智能问答和代码补全功能，支持多种编程语言，帮助开发者提高编程效率。',
+      description:
+        'CodeGeeX是一款强大的AI编程助手，提供智能问答和代码补全功能，支持多种编程语言，帮助开发者提高编程效率。',
       displayName: 'CodeGeeX-4',
       enabled: true,
       functionCall: false,

src/libs/agent-runtime/openai/__snapshots__/index.test.ts.snap

@@ -3,7 +3,7 @@
 exports[`LobeOpenAI > models > should get models 1`] = `
 [
   {
-    "description": "GPT 3.5 Turbo，适用于各种文本生成和理解任务",
+    "description": "GPT 3.5 Turbo，适用于各种文本生成和理解任务，Currently points to gpt-3.5-turbo-0125",
     "displayName": "GPT-3.5 Turbo",
     "functionCall": true,
     "id": "gpt-3.5-turbo",
@@ -17,7 +17,7 @@ exports[`LobeOpenAI > models > should get models 1`] = `
     "tokens": 16385,
   },
   {
-    "displayName": "GPT-3.5 Turbo 16K (0613)",
+    "description": "Currently points to gpt-3.5-turbo-16k-0613",
     "id": "gpt-3.5-turbo-16k-0613",
     "legacy": true,
     "tokens": 16385,

src/libs/agent-runtime/openrouter/__snapshots__/index.test.ts.snap

@@ -59,7 +59,7 @@ Updated by OpenAI to point to the [latest version of GPT-3.5](/models?q=openai/g
 
 Usage of Gemma is subject to Google's [Gemma Terms of Use](https://ai.google.dev/gemma/terms).",
     "displayName": "Google: Gemma 7B",
-    "enabled": true,
+    "enabled": false,
     "functionCall": false,
     "id": "google/gemma-7b-it",
     "maxTokens": undefined,

src/libs/agent-runtime/togetherai/__snapshots__/index.test.ts.snap

@@ -135,7 +135,7 @@ exports[`LobeTogetherAI > models > should get models 1`] = `
   {
     "description": "Qwen1.5 is the beta version of Qwen2, a transformer-based decoder-only language model pretrained on a large amount of data. In comparison with the previous released Qwen.",
     "displayName": "Qwen 1.5 Chat (14B)",
-    "enabled": true,
+    "enabled": false,
     "functionCall": false,
     "id": "Qwen/Qwen1.5-14B-Chat",
     "maxOutput": 32768,
@@ -145,7 +145,7 @@ exports[`LobeTogetherAI > models > should get models 1`] = `
   {
     "description": "Qwen1.5 is the beta version of Qwen2, a transformer-based decoder-only language model pretrained on a large amount of data. In comparison with the previous released Qwen.",
     "displayName": "Qwen 1.5 Chat (32B)",
-    "enabled": false,
+    "enabled": true,
     "functionCall": false,
     "id": "Qwen/Qwen1.5-32B-Chat",
     "maxOutput": 32768,
@@ -335,7 +335,7 @@ exports[`LobeTogetherAI > models > should get models 1`] = `
   {
     "description": "Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models.",
     "displayName": "Gemma Instruct (7B)",
-    "enabled": true,
+    "enabled": false,
     "functionCall": false,
     "id": "google/gemma-7b-it",
     "maxOutput": 8192,