@lobehub/chat 1.16.7 → 1.16.9

package/src/config/modelProviders/bedrock.ts

@@ -1,123 +1,182 @@
 import { ModelProviderCard } from '@/types/llm';
 
-// ref https://docs.aws.amazon.com/bedrock/latest/userguide/model-ids.html
-// ref https://docs.aws.amazon.com/bedrock/latest/userguide/conversation-inference.html
-// ref https://us-east-1.console.aws.amazon.com/bedrock/home?region=us-east-1#/models
-// ref https://us-west-2.console.aws.amazon.com/bedrock/home?region=us-west-2#/models
+// ref :https://docs.aws.amazon.com/bedrock/latest/userguide/conversation-inference.html
+// ref :https://us-east-1.console.aws.amazon.com/bedrock/home?region=us-east-1#/models
+// ref :https://us-west-2.console.aws.amazon.com/bedrock/home?region=us-west-2#/models
 const Bedrock: ModelProviderCard = {
   chatModels: [
-/*
+    /*
     // TODO: Not support for now
     {
-      description: 'Amazon Titan Text Lite is a light weight efficient model ideal for fine-tuning for English-language tasks, including like summarization and copywriting, where customers want a smaller, more cost-effective model that is also highly customizable.',
+      description: '亚马逊 Titan Text Lite 是一款轻量级高效模型，非常适合对英语任务进行微调，包括总结和文案编写等，客户希望有一个更小、更经济的模型，同时也非常可定制。',
       displayName: 'Titan Text G1 - Lite',
       id: 'amazon.titan-text-lite-v1',
       tokens: 4000,
     },
     {
-      description: 'Amazon Titan Text Express has a context length of up to 8,000 tokens, making it well-suited for a wide range of advanced, general language tasks such as open-ended text generation and conversational chat, as well as support within Retrieval Augmented Generation (RAG). At launch, the model is optimized for English, with multilingual support for more than 100 additional languages available in preview.',
+      description: '亚马逊 Titan Text Express 的上下文长度可达 8,000 个标记，非常适合广泛的高级通用语言任务，如开放式文本生成和对话聊天，以及在检索增强生成 (RAG) 中的支持。在推出时，该模型针对英语进行了优化，预览版还支持其他 100 多种语言。',
       displayName: 'Titan Text G1 - Express',
       id: 'amazon.titan-text-express-v1',
       tokens: 8000,
     },
     {
-      description: 'Titan Text Premier is a powerful and advanced model within the Titan Text family, designed to deliver superior performance across a wide range of enterprise applications. With its cutting-edge capabilities, it offers enhanced accuracy and exceptional results, making it an excellent choice for organizations seeking top-notch text processing solutions.',
+      description: 'Titan Text Premier 是 Titan Text 系列中一款强大的先进模型，旨在为广泛的企业应用提供卓越的性能。凭借其尖端能力，它提供了更高的准确性和卓越的结果，是寻求一流文本处理解决方案的组织的绝佳选择。',
       displayName: 'Titan Text G1 - Premier',
       id: 'amazon.titan-text-premier-v1:0',
       tokens: 32_000,
     },
 */
     {
-      description: 'Claude 3.5 Sonnet raises the industry bar for intelligence, outperforming competitor models and Claude 3 Opus on a wide range of evaluations, with the speed and cost of our mid-tier model, Claude 3 Sonnet.',
+      description:
+        'Claude 3.5 Sonnet 提升了行业标准，性能超过竞争对手模型和 Claude 3 Opus，在广泛的评估中表现出色，同时具有我们中等层级模型的速度和成本。',
       displayName: 'Claude 3.5 Sonnet',
       enabled: true,
       functionCall: true,
       id: 'anthropic.claude-3-5-sonnet-20240620-v1:0',
+      pricing: {
+        input: 3,
+        output: 15,
+      },
       tokens: 200_000,
       vision: true,
     },
     {
-      description: 'Claude 3 Sonnet by Anthropic strikes the ideal balance between intelligence and speed—particularly for enterprise workloads. It offers maximum utility at a lower price than competitors, and is engineered to be the dependable, high-endurance workhorse for scaled AI deployments. Claude 3 Sonnet can process images and return text outputs, and features a 200K context window.',
+      description:
+        'Anthropic 的 Claude 3 Sonnet 在智能和速度之间达到了理想的平衡——特别适合企业工作负载。它以低于竞争对手的价格提供最大的效用，并被设计成为可靠的、高耐用的主力机，适用于规模化的 AI 部署。Claude 3 Sonnet 可以处理图像并返回文本输出，具有 200K 的上下文窗口。',
       displayName: 'Claude 3 Sonnet',
       enabled: true,
       functionCall: true,
       id: 'anthropic.claude-3-sonnet-20240229-v1:0',
+      pricing: {
+        input: 3,
+        output: 15,
+      },
       tokens: 200_000,
       vision: true,
     },
     {
-      description: 'Claude 3 Opus is Anthropic most powerful AI model, with state-of-the-art performance on highly complex tasks. It can navigate open-ended prompts and sight-unseen scenarios with remarkable fluency and human-like understanding. Claude 3 Opus shows us the frontier of what’s possible with generative AI. Claude 3 Opus can process images and return text outputs, and features a 200K context window.',
+      description:
+        'Claude 3 Opus 是 Anthropic 最强大的 AI 模型，具有在高度复杂任务上的最先进性能。它可以处理开放式提示和未见过的场景，具有出色的流畅性和类人的理解能力。Claude 3 Opus 展示了生成 AI 可能性的前沿。Claude 3 Opus 可以处理图像并返回文本输出，具有 200K 的上下文窗口。',
       displayName: 'Claude 3 Opus',
       enabled: true,
       functionCall: true,
       id: 'anthropic.claude-3-opus-20240229-v1:0',
+      pricing: {
+        input: 15,
+        output: 75,
+      },
       tokens: 200_000,
       vision: true,
     },
     {
-      description: 'Claude 3 Haiku is Anthropic fastest, most compact model for near-instant responsiveness. It answers simple queries and requests with speed. Customers will be able to build seamless AI experiences that mimic human interactions. Claude 3 Haiku can process images and return text outputs, and features a 200K context window.',
+      description:
+        'Claude 3 Haiku 是 Anthropic 最快、最紧凑的模型，提供近乎即时的响应速度。它可以快速回答简单的查询和请求。客户将能够构建模仿人类互动的无缝 AI 体验。Claude 3 Haiku 可以处理图像并返回文本输出，具有 200K 的上下文窗口。',
       displayName: 'Claude 3 Haiku',
       enabled: true,
       functionCall: true,
       id: 'anthropic.claude-3-haiku-20240307-v1:0',
+      pricing: {
+        input: 0.25,
+        output: 1.25,
+      },
       tokens: 200_000,
       vision: true,
     },
     {
-      description: 'An update to Claude 2 that features double the context window, plus improvements across reliability, hallucination rates, and evidence-based accuracy in long document and RAG contexts.',
+      description:
+        'Claude 2 的更新版，具有双倍的上下文窗口，以及在长文档和 RAG 上下文中的可靠性、幻觉率和基于证据的准确性的改进。',
       displayName: 'Claude 2.1',
       id: 'anthropic.claude-v2:1',
+      pricing: {
+        input: 8,
+        output: 24,
+      },
       tokens: 200_000,
     },
     {
-      description: 'Anthropic highly capable model across a wide range of tasks from sophisticated dialogue and creative content generation to detailed instruction following.',
+      description:
+        'Anthropic 在从复杂对话和创意内容生成到详细指令跟随的广泛任务中都表现出高度能力的模型。',
       displayName: 'Claude 2.0',
       id: 'anthropic.claude-v2',
+      pricing: {
+        input: 8,
+        output: 24,
+      },
       tokens: 100_000,
     },
     {
-      description: 'A fast, affordable yet still very capable model, which can handle a range of tasks including casual dialogue, text analysis, summarization, and document question-answering.',
+      description:
+        '一款快速、经济且仍然非常有能力的模型，可以处理包括日常对话、文本分析、总结和文档问答在内的一系列任务。',
       displayName: 'Claude Instant',
       id: 'anthropic.claude-instant-v1',
+      pricing: {
+        input: 0.8,
+        output: 2.4,
+      },
       tokens: 100_000,
     },
     {
-      description: 'An update to Meta Llama 3 8B Instruct that includes an expanded 128K context length, multilinguality and improved reasoning capabilities. The Llama 3.1 offering of multilingual large language models (LLMs) is a collection of pretrained and instruction-tuned generative models in 8B, 70B and 405B sizes (text in/text out). The Llama 3.1 instruction-tuned text only models (8B, 70B, 405B) are optimized for multilingual dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Llama 3.1 is intended for commercial and research use in multiple languages. Instruction tuned text only models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. The Llama 3.1 models also support the ability to leverage the outputs of its models to improve other models including synthetic data generation and distillation. Llama 3.1 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.',
+      description:
+        'Meta Llama 3.1 8B Instruct 的更新版，包括扩展的 128K 上下文长度、多语言性和改进的推理能力。Llama 3.1 提供的多语言大型语言模型 (LLMs) 是一组预训练的、指令调整的生成模型，包括 8B、70B 和 405B 大小 (文本输入/输出)。Llama 3.1 指令调整的文本模型 (8B、70B、405B) 专为多语言对话用例进行了优化，并在常见的行业基准测试中超过了许多可用的开源聊天模型。Llama 3.1 旨在用于多种语言的商业和研究用途。指令调整的文本模型适用于类似助手的聊天，而预训练模型可以适应各种自然语言生成任务。Llama 3.1 模型还支持利用其模型的输出来改进其他模型，包括合成数据生成和精炼。Llama 3.1 是使用优化的变压器架构的自回归语言模型。调整版本使用监督微调 (SFT) 和带有人类反馈的强化学习 (RLHF) 来符合人类对帮助性和安全性的偏好。',
       displayName: 'Llama 3.1 8B Instruct',
       enabled: true,
       functionCall: true,
       id: 'meta.llama3-1-8b-instruct-v1:0',
+      pricing: {
+        input: 0.22,
+        output: 0.22,
+      },
       tokens: 128_000,
     },
     {
-      description: 'An update to Meta Llama 3 70B Instruct that includes an expanded 128K context length, multilinguality and improved reasoning capabilities. The Llama 3.1 offering of multilingual large language models (LLMs) is a collection of pretrained and instruction-tuned generative models in 8B, 70B and 405B sizes (text in/text out). The Llama 3.1 instruction-tuned text only models (8B, 70B, 405B) are optimized for multilingual dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Llama 3.1 is intended for commercial and research use in multiple languages. Instruction tuned text only models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. The Llama 3.1 models also support the ability to leverage the outputs of its models to improve other models including synthetic data generation and distillation. Llama 3.1 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.',
+      description:
+        'Meta Llama 3.1 70B Instruct 的更新版，包括扩展的 128K 上下文长度、多语言性和改进的推理能力。Llama 3.1 提供的多语言大型语言模型 (LLMs) 是一组预训练的、指令调整的生成模型，包括 8B、70B 和 405B 大小 (文本输入/输出)。Llama 3.1 指令调整的文本模型 (8B、70B、405B) 专为多语言对话用例进行了优化，并在常见的行业基准测试中超过了许多可用的开源聊天模型。Llama 3.1 旨在用于多种语言的商业和研究用途。指令调整的文本模型适用于类似助手的聊天，而预训练模型可以适应各种自然语言生成任务。Llama 3.1 模型还支持利用其模型的输出来改进其他模型，包括合成数据生成和精炼。Llama 3.1 是使用优化的变压器架构的自回归语言模型。调整版本使用监督微调 (SFT) 和带有人类反馈的强化学习 (RLHF) 来符合人类对帮助性和安全性的偏好。',
       displayName: 'Llama 3.1 70B Instruct',
       enabled: true,
       functionCall: true,
       id: 'meta.llama3-1-70b-instruct-v1:0',
+      pricing: {
+        input: 0.99,
+        output: 0.99,
+      },
       tokens: 128_000,
     },
     {
-      description: 'Meta Llama 3.1 405B Instruct is the largest and most powerful of the Llama 3.1 Instruct models that is a highly advanced model for conversational inference and reasoning, synthetic data generation, and a base to do specialized continual pre-training or fine-tuning on a specific domain. The Llama 3.1 offering of multilingual large language models (LLMs) is a collection of pretrained and instruction-tuned generative models in 8B, 70B and 405B sizes (text in/text out). The Llama 3.1 instruction-tuned text only models (8B, 70B, 405B) are optimized for multilingual dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Llama 3.1 is intended for commercial and research use in multiple languages. Instruction tuned text only models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. The Llama 3.1 models also support the ability to leverage the outputs of its models to improve other models including synthetic data generation and distillation. Llama 3.1 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.',
+      description:
+        'Meta Llama 3.1 405B Instruct 是 Llama 3.1 Instruct 模型中最大、最强大的模型，是一款高度先进的对话推理和合成数据生成模型，也可以用作在特定领域进行专业持续预训练或微调的基础。Llama 3.1 提供的多语言大型语言模型 (LLMs) 是一组预训练的、指令调整的生成模型，包括 8B、70B 和 405B 大小 (文本输入/输出)。Llama 3.1 指令调整的文本模型 (8B、70B、405B) 专为多语言对话用例进行了优化，并在常见的行业基准测试中超过了许多可用的开源聊天模型。Llama 3.1 旨在用于多种语言的商业和研究用途。指令调整的文本模型适用于类似助手的聊天，而预训练模型可以适应各种自然语言生成任务。Llama 3.1 模型还支持利用其模型的输出来改进其他模型，包括合成数据生成和精炼。Llama 3.1 是使用优化的变压器架构的自回归语言模型。调整版本使用监督微调 (SFT) 和带有人类反馈的强化学习 (RLHF) 来符合人类对帮助性和安全性的偏好。',
       displayName: 'Llama 3.1 405B Instruct',
       enabled: true,
       functionCall: true,
       id: 'meta.llama3-1-405b-instruct-v1:0',
+      pricing: {
+        input: 5.32,
+        output: 16,
+      },
       tokens: 128_000,
     },
     {
-      description: 'Meta Llama 3 is an accessible, open large language model (LLM) designed for developers, researchers, and businesses to build, experiment, and responsibly scale their generative AI ideas. Part of a foundational system, it serves as a bedrock for innovation in the global community. Ideal for limited computational power and resources, edge devices, and faster training times.',
+      description:
+        'Meta Llama 3 是一款面向开发者、研究人员和企业的开放大型语言模型 (LLM)，旨在帮助他们构建、实验并负责任地扩展他们的生成 AI 想法。作为全球社区创新的基础系统的一部分，它非常适合计算能力和资源有限、边缘设备和更快的训练时间。',
       displayName: 'Llama 3 8B Instruct',
       id: 'meta.llama3-8b-instruct-v1:0',
+      pricing: {
+        input: 0.3,
+        output: 0.6,
+      },
       tokens: 8000,
     },
     {
-      description: 'Meta Llama 3 is an accessible, open large language model (LLM) designed for developers, researchers, and businesses to build, experiment, and responsibly scale their generative AI ideas. Part of a foundational system, it serves as a bedrock for innovation in the global community. Ideal for content creation, conversational AI, language understanding, R&D, and Enterprise applications.',
+      description:
+        'Meta Llama 3 是一款面向开发者、研究人员和企业的开放大型语言模型 (LLM)，旨在帮助他们构建、实验并负责任地扩展他们的生成 AI 想法。作为全球社区创新的基础系统的一部分，它非常适合内容创建、对话 AI、语言理解、研发和企业应用。',
       displayName: 'Llama 3 70B Instruct',
       id: 'meta.llama3-70b-instruct-v1:0',
+      pricing: {
+        input: 2.65,
+        output: 3.5,
+      },
       tokens: 8000,
     },
-/*
+    /*
     // TODO: Not support for now
     {
       description: 'A 7B dense Transformer, fast-deployed and easily customisable. Small, yet powerful for a variety of use cases. Supports English and code, and a 32k context window.',
@@ -157,7 +216,7 @@ const Bedrock: ModelProviderCard = {
       tokens: 32_000,
     },
 */
-/*
+    /*
     // TODO: Not support for now
     {
       description: 'Command R+ is a highly performant generative language model optimized for large scale production workloads.',
@@ -176,7 +235,7 @@ const Bedrock: ModelProviderCard = {
       tokens: 128_000,
     },
 */
-/*
+    /*
     // Cohere Command (Text) and AI21 Labs Jurassic-2 (Text) don't support chat with the Converse API
     {
       description: 'Command is Cohere flagship text generation model. It is trained to follow user commands and to be instantly useful in practical business applications.',
@@ -191,7 +250,7 @@ const Bedrock: ModelProviderCard = {
       tokens: 4000,
     },
 */
-/*
+    /*
     // TODO: Not support for now
     {
       description: 'The latest Foundation Model from AI21 Labs, Jamba-Instruct offers an impressive 256K context window and delivers the best value per price on core text generation, summarization, and question answering tasks for the enterprise.',
@@ -200,7 +259,7 @@ const Bedrock: ModelProviderCard = {
       tokens: 256_000,
     },
 */
-/*
+    /*
     // Cohere Command (Text) and AI21 Labs Jurassic-2 (Text) don't support chat with the Converse API
     {
       description: 'Jurassic-2 Mid is less powerful than Ultra, yet carefully designed to strike the right balance between exceptional quality and affordability. Jurassic-2 Mid can be applied to any language comprehension or generation task including question answering, summarization, long-form copy generation, advanced information extraction and many others.',
@@ -217,8 +276,12 @@ const Bedrock: ModelProviderCard = {
 */
   ],
   checkModel: 'anthropic.claude-instant-v1',
+  description:
+    'Bedrock 是亚马逊 AWS 提供的一项服务，专注于为企业提供先进的 AI 语言模型和视觉模型。其模型家族包括 Anthropic 的 Claude 系列、Meta 的 Llama 3.1 系列等，涵盖从轻量级到高性能的多种选择，支持文本生成、对话、图像处理等多种任务，适用于不同规模和需求的企业应用。',
   id: 'bedrock',
+  modelsUrl: 'https://docs.aws.amazon.com/bedrock/latest/userguide/model-ids.html',
   name: 'Bedrock',
+  url: 'https://docs.aws.amazon.com/bedrock/latest/userguide/what-is-bedrock.html',
 };
 
 export default Bedrock;

package/src/config/modelProviders/deepseek.ts

@@ -1,21 +1,32 @@
 import { ModelProviderCard } from '@/types/llm';
 
-// ref https://platform.deepseek.com/api-docs/pricing
+// ref: https://platform.deepseek.com/api-docs/pricing
 const DeepSeek: ModelProviderCard = {
   chatModels: [
     {
-      description: '擅长通用对话任务',
-      displayName: 'DeepSeek V2.5',
+      description:
+        '融合通用与代码能力的全新开源模型, 不仅保留了原有 Chat 模型的通用对话能力和 Coder 模型的强大代码处理能力，还更好地对齐了人类偏好。此外，DeepSeek-V2.5 在写作任务、指令跟随等多个方面也实现了大幅提升。',
+      displayName: 'DeepSeek-V2.5',
       enabled: true,
       functionCall: true,
       id: 'deepseek-chat',
+      pricing: {
+        cachedInput: 0.014,
+        input: 0.14,
+        output: 0.28,
+      },
+      releasedAt: '2024-09-05',
       tokens: 128_000,
     },
   ],
   checkModel: 'deepseek-chat',
+  description:
+    'DeepSeek 是一家专注于人工智能技术研究和应用的公司，其最新模型 DeepSeek-V2.5 融合了通用对话和代码处理能力，并在人类偏好对齐、写作任务和指令跟随等方面实现了显著提升。',
   id: 'deepseek',
   modelList: { showModelFetcher: true },
+  modelsUrl: 'https://platform.deepseek.com/api-docs/zh-cn/quick_start/pricing',
   name: 'DeepSeek',
+  url: 'https://deepseek.com',
 };
 
 export default DeepSeek;

package/src/config/modelProviders/fireworksai.ts

@@ -1,11 +1,12 @@
 import { ModelProviderCard } from '@/types/llm';
 
-// ref https://fireworks.ai/models?show=Serverless
-// ref https://fireworks.ai/pricing
+// ref: https://fireworks.ai/models?show=Serverless
+// ref: https://fireworks.ai/pricing
 const FireworksAI: ModelProviderCard = {
   chatModels: [
     {
-      description: 'Fireworks latest and most performant function-calling model. Firefunction-v2 is based on Llama-3 and trained to excel at function-calling as well as chat and instruction-following. See blog post for more details https://fireworks.ai/blog/firefunction-v2-launch-post',
+      description:
+        'Fireworks 公司最新推出的 Firefunction-v2 是一款性能卓越的函数调用模型，基于 Llama-3 开发，并通过大量优化，特别适用于函数调用、对话及指令跟随等场景。',
       displayName: 'Firefunction V2',
       enabled: true,
       functionCall: true,
@@ -13,14 +14,15 @@ const FireworksAI: ModelProviderCard = {
       tokens: 8192,
     },
     {
-      description: 'Fireworks open-source function calling model.',
+      description: 'Fireworks 开源函数调用模型，提供卓越的指令执行能力和开放可定制的特性。',
       displayName: 'Firefunction V1',
       functionCall: true,
       id: 'accounts/fireworks/models/firefunction-v1',
       tokens: 32_768,
     },
     {
-      description: 'Vision-language model allowing both image and text as inputs (single image is recommended), trained on OSS model generated training data and open sourced on huggingface at fireworks-ai/FireLLaVA-13b',
+      description:
+        'fireworks-ai/FireLLaVA-13b 是一款视觉语言模型，可以同时接收图像和文本输入，经过高质量数据训练，适合多模态任务。',
       displayName: 'FireLLaVA-13B',
       enabled: true,
       functionCall: false,
@@ -29,6 +31,8 @@ const FireworksAI: ModelProviderCard = {
       vision: true,
     },
     {
+      description:
+        'Llama 3.1 8B 指令模型，专为多语言对话优化，能够在常见行业基准上超越多数开源及闭源模型。',
       displayName: 'Llama 3.1 8B Instruct',
       enabled: true,
       functionCall: false,
@@ -36,6 +40,8 @@ const FireworksAI: ModelProviderCard = {
       tokens: 131_072,
     },
     {
+      description:
+        'Llama 3.1 70B 指令模型，提供卓越的自然语言理解和生成能力，是对话及分析任务的理想选择。',
       displayName: 'Llama 3.1 70B Instruct',
       enabled: true,
       functionCall: false,
@@ -43,6 +49,8 @@ const FireworksAI: ModelProviderCard = {
       tokens: 131_072,
     },
     {
+      description:
+        'Llama 3.1 405B 指令模型，具备超大规模参数，适合复杂任务和高负载场景下的指令跟随。',
       displayName: 'Llama 3.1 405B Instruct',
       enabled: true,
       functionCall: false,
@@ -50,30 +58,38 @@ const FireworksAI: ModelProviderCard = {
       tokens: 131_072,
     },
     {
+      description: 'Llama 3 8B 指令模型，优化用于对话及多语言任务，表现卓越且高效。',
       displayName: 'Llama 3 8B Instruct',
       functionCall: false,
       id: 'accounts/fireworks/models/llama-v3-8b-instruct',
       tokens: 8192,
     },
     {
+      description: 'Llama 3 70B 指令模型，专为多语言对话和自然语言理解优化，性能优于多数竞争模型。',
       displayName: 'Llama 3 70B Instruct',
       functionCall: false,
       id: 'accounts/fireworks/models/llama-v3-70b-instruct',
       tokens: 8192,
     },
     {
+      description:
+        'Llama 3 8B 指令模型（HF 版本），与官方实现结果一致，具备高度一致性和跨平台兼容性。',
       displayName: 'Llama 3 8B Instruct (HF version)',
       functionCall: false,
       id: 'accounts/fireworks/models/llama-v3-8b-instruct-hf',
       tokens: 8192,
     },
     {
+      description:
+        'Llama 3 70B 指令模型（HF 版本），与官方实现结果保持一致，适合高质量的指令跟随任务。',
       displayName: 'Llama 3 70B Instruct (HF version)',
       functionCall: false,
       id: 'accounts/fireworks/models/llama-v3-70b-instruct-hf',
       tokens: 8192,
     },
     {
+      description:
+        'Gemma 2 9B 指令模型，基于之前的Google技术，适合回答问题、总结和推理等多种文本生成任务。',
       displayName: 'Gemma 2 9B Instruct',
       enabled: true,
       functionCall: false,
@@ -81,6 +97,7 @@ const FireworksAI: ModelProviderCard = {
       tokens: 8192,
     },
     {
+      description: 'Mixtral MoE 8x7B 指令模型，多专家架构提供高效的指令跟随及执行。',
       displayName: 'Mixtral MoE 8x7B Instruct',
       enabled: true,
       functionCall: false,
@@ -88,6 +105,8 @@ const FireworksAI: ModelProviderCard = {
       tokens: 32_768,
     },
     {
+      description:
+        'Mixtral MoE 8x22B 指令模型，大规模参数和多专家架构，全方位支持复杂任务的高效处理。',
       displayName: 'Mixtral MoE 8x22B Instruct',
       enabled: true,
       functionCall: false,
@@ -95,12 +114,16 @@ const FireworksAI: ModelProviderCard = {
       tokens: 65_536,
     },
     {
+      description:
+        'Mixtral MoE 8x7B 指令模型（HF 版本），性能与官方实现一致，适合多种高效任务场景。',
       displayName: 'Mixtral MoE 8x7B Instruct (HF version)',
       functionCall: false,
       id: 'accounts/fireworks/models/mixtral-8x7b-instruct-hf',
       tokens: 32_768,
     },
     {
+      description:
+        'Phi 3 Vision 指令模型，轻量级多模态模型，能够处理复杂的视觉和文本信息，具备较强的推理能力。',
       displayName: 'Phi 3 Vision Instruct',
       enabled: true,
       functionCall: false,
@@ -109,6 +132,7 @@ const FireworksAI: ModelProviderCard = {
       vision: true,
     },
     {
+      description: 'Yi-Large 模型，具备卓越的多语言处理能力，可用于各类语言生成和理解任务。',
       displayName: 'Yi-Large',
       enabled: true,
       functionCall: false,
@@ -116,18 +140,22 @@ const FireworksAI: ModelProviderCard = {
       tokens: 32_768,
     },
     {
+      description: 'StarCoder 7B 模型，针对80多种编程语言训练，拥有出色的编程填充能力和语境理解。',
       displayName: 'StarCoder 7B',
       functionCall: false,
       id: 'accounts/fireworks/models/starcoder-7b',
       tokens: 8192,
     },
     {
+      description:
+        'StarCoder 15.5B 模型，支持高级编程任务，多语言能力增强，适合复杂代码生成和理解。',
       displayName: 'StarCoder 15.5B',
       functionCall: false,
       id: 'accounts/fireworks/models/starcoder-16b',
       tokens: 8192,
     },
     {
+      description: 'MythoMax L2 13B 模型，结合新颖的合并技术，擅长叙事和角色扮演。',
       displayName: 'MythoMax L2 13b',
       functionCall: false,
       id: 'accounts/fireworks/models/mythomax-l2-13b',
@@ -135,9 +163,13 @@ const FireworksAI: ModelProviderCard = {
     },
   ],
   checkModel: 'accounts/fireworks/models/firefunction-v2',
+  description:
+    'Fireworks AI 是一家领先的高级语言模型服务商，专注于功能调用和多模态处理。其最新模型 Firefunction V2 基于 Llama-3，优化用于函数调用、对话及指令跟随。视觉语言模型 FireLLaVA-13B 支持图像和文本混合输入。其他 notable 模型包括 Llama 系列和 Mixtral 系列，提供高效的多语言指令跟随与生成支持。',
   id: 'fireworksai',
   modelList: { showModelFetcher: true },
+  modelsUrl: 'https://fireworks.ai/models?show=Serverless',
   name: 'Fireworks AI',
+  url: 'https://fireworks.ai',
 };
 
 export default FireworksAI;

package/src/config/modelProviders/google.ts

@@ -1,40 +1,52 @@
 import { ModelProviderCard } from '@/types/llm';
 
-// ref https://ai.google.dev/models/gemini
-// api https://ai.google.dev/api/rest/v1beta/models/list
+// ref: https://ai.google.dev/gemini-api/docs/models/gemini
 const Google: ModelProviderCard = {
   chatModels: [
     {
-      description: 'Fast and versatile multimodal model for scaling across diverse tasks',
+      description:
+        'Gemini 1.5 Flash 是Google最新的多模态AI模型，具备快速处理能力，支持文本、图像和视频输入，适用于多种任务的高效扩展。',
       displayName: 'Gemini 1.5 Flash',
       enabled: true,
       functionCall: true,
       id: 'gemini-1.5-flash-latest',
       maxOutput: 8192,
+      pricing: {
+        cachedInput: 0.018_75,
+        input: 0.075,
+        output: 0.3,
+      },
       tokens: 1_048_576 + 8192,
       vision: true,
     },
     {
-      description: 'Fast and versatile multimodal model for scaling across diverse tasks',
+      description: 'Gemini 1.5 Flash 0827 提供了优化后的多模态处理能力，适用多种复杂任务场景。',
       displayName: 'Gemini 1.5 Flash 0827',
       enabled: true,
       functionCall: true,
       id: 'gemini-1.5-flash-exp-0827',
       maxOutput: 8192,
+      pricing: {
+        cachedInput: 0.018_75,
+        input: 0.075,
+        output: 0.3,
+      },
+      releasedAt: '2024-08-27',
       tokens: 1_048_576 + 8192,
       vision: true,
     },
     {
-      description: 'Fast and versatile multimodal model for scaling across diverse tasks',
+      description: 'Gemini 1.5 Flash 8B 0827 专为处理大规模任务场景设计，提供无与伦比的处理速度。',
       displayName: 'Gemini 1.5 Flash 8B 0827',
       functionCall: true,
       id: 'gemini-1.5-flash-8b-exp-0827',
       maxOutput: 8192,
+      releasedAt: '2024-08-27',
       tokens: 1_048_576 + 8192,
       vision: true,
     },
     {
-      description: 'Fast and versatile multimodal model for scaling across diverse tasks',
+      description: 'Gemini 1.5 Flash 001 是一款高效的多模态模型，支持广泛应用的扩展。',
       displayName: 'Gemini 1.5 Flash 001',
       functionCall: true,
       id: 'gemini-1.5-flash-001',
@@ -43,71 +55,112 @@ const Google: ModelProviderCard = {
       vision: true,
     },
     {
-      description: 'Mid-size multimodal model that supports up to 2 million tokens',
+      description:
+        'Gemini 1.5 Pro 支持高达200万个tokens，是中型多模态模型的理想选择，适用于复杂任务的多方面支持。',
       displayName: 'Gemini 1.5 Pro',
       enabled: true,
       functionCall: true,
       id: 'gemini-1.5-pro-latest',
       maxOutput: 8192,
+      pricing: {
+        cachedInput: 0.875,
+        input: 3.5,
+        output: 10.5,
+      },
+      releasedAt: '2024-02-15',
       tokens: 2_097_152 + 8192,
       vision: true,
     },
     {
-      description: 'Mid-size multimodal model that supports up to 2 million tokens',
+      description: 'Gemini 1.5 Pro 0827 结合最新优化技术，带来更高效的多模态数据处理能力。',
       displayName: 'Gemini 1.5 Pro 0827',
       enabled: true,
       functionCall: true,
       id: 'gemini-1.5-pro-exp-0827',
       maxOutput: 8192,
+      pricing: {
+        cachedInput: 0.875,
+        input: 3.5,
+        output: 10.5,
+      },
+      releasedAt: '2024-08-27',
       tokens: 2_097_152 + 8192,
       vision: true,
     },
     {
-      description: 'Mid-size multimodal model that supports up to 2 million tokens',
+      description: 'Gemini 1.5 Pro 0801 提供出色的多模态处理能力，为应用开发带来更大灵活性。',
       displayName: 'Gemini 1.5 Pro 0801',
       functionCall: true,
       id: 'gemini-1.5-pro-exp-0801',
       maxOutput: 8192,
+      pricing: {
+        cachedInput: 0.875,
+        input: 3.5,
+        output: 10.5,
+      },
+      releasedAt: '2024-08-01',
       tokens: 2_097_152 + 8192,
       vision: true,
     },
     {
-      description: 'Mid-size multimodal model that supports up to 2 million tokens',
+      description: 'Gemini 1.5 Pro 001 是可扩展的多模态AI解决方案，支持广泛的复杂任务。',
       displayName: 'Gemini 1.5 Pro 001',
       functionCall: true,
       id: 'gemini-1.5-pro-001',
       maxOutput: 8192,
+      pricing: {
+        cachedInput: 0.875,
+        input: 3.5,
+        output: 10.5,
+      },
+      releasedAt: '2024-02-15',
       tokens: 2_097_152 + 8192,
       vision: true,
     },
     {
-      description:
-        'The best model for scaling across a wide range of tasks. This is the latest model.',
+      description: 'Gemini 1.0 Pro 是Google的高性能AI模型，专为广泛任务扩展而设计。',
       displayName: 'Gemini 1.0 Pro',
       id: 'gemini-1.0-pro-latest',
       maxOutput: 2048,
+      pricing: {
+        input: 0.5,
+        output: 1.5,
+      },
+      releasedAt: '2023-12-06',
       tokens: 30_720 + 2048,
     },
     {
       description:
-        'The best model for scaling across a wide range of tasks. This is a stable model that supports tuning.',
+        'Gemini 1.0 Pro 001 (Tuning) 提供稳定并可调优的性能，是复杂任务解决方案的理想选择。',
       displayName: 'Gemini 1.0 Pro 001 (Tuning)',
       functionCall: true,
       id: 'gemini-1.0-pro-001',
       maxOutput: 2048,
+      pricing: {
+        input: 0.5,
+        output: 1.5,
+      },
+      releasedAt: '2023-12-06',
       tokens: 30_720 + 2048,
     },
     {
-      description:
-        'The best model for scaling across a wide range of tasks. Released April 9, 2024.',
+      description: 'Gemini 1.0 Pro 002 (Tuning) 提供出色的多模态支持，专注于复杂任务的有效解决。',
       displayName: 'Gemini 1.0 Pro 002 (Tuning)',
       id: 'gemini-1.0-pro-002',
       maxOutput: 2048,
+      pricing: {
+        input: 0.5,
+        output: 1.5,
+      },
+      releasedAt: '2023-12-06',
       tokens: 30_720 + 2048,
     },
   ],
   checkModel: 'gemini-1.5-flash-latest',
+  description:
+    'Google 的 Gemini 系列是其最先进、通用的 A I模型，由 Google DeepMind 打造，专为多模态设计，支持文本、代码、图像、音频和视频的无缝理解与处理。适用于从数据中心到移动设备的多种环境，极大提升了AI模型的效率与应用广泛性。',
   id: 'google',
+  modelsUrl: 'https://ai.google.dev/gemini-api/docs/models/gemini',
   name: 'Google',
   proxyUrl: {
     placeholder: 'https://generativelanguage.googleapis.com',
@@ -116,6 +169,7 @@ const Google: ModelProviderCard = {
     speed: 2,
     text: true,
   },
+  url: 'https://ai.google.dev',
 };
 
 export default Google;