上游供应网络

国内知名 AI API 聚合平台，模型覆盖全面

领先的AI研究公司，提供GPT-4、GPT-3.5、DALL-E和Whisper API

开源一站式中转服务，支持Claude、OpenAI、Gemini、Droid统一接入和拼车

Google 的 AI 平台，提供 Gemini 模型。多模态能力，具有竞争力的价格和与 Google Cloud 的集成。

谷歌的Vertex AI提供Gemini模型和其他LLM的访问，与Google Cloud服务无缝集成。支持多模态、推理和企业级部署，具有强大的安全和合规功能。适合构建可扩展AI应用的开发者。

Hugging Face提供数千个开源模型的无服务器推理API，包括Llama、Mistral和自定义微调。支持文本、视觉、音频，具有简单部署和扩展。适合原型设计和生产。

阿里云的通义千问（Qwen）系列提供多语言LLM，在中文和英文方面表现强劲。通过API提供，具有企业功能和大上下文支持。

NVIDIA NIM provides optimized inference microservices for LLMs on NVIDIA GPUs.

Cloudflare Workers AI offers serverless inference for LLMs at the edge with low latency.

Lightning AI provides studio and inference for PyTorch-based LLM development and deployment.

Clarifai provides AI platform with LLM inference and multimodal capabilities.

小米通过其生态提供AI模型和API访问。

xAI 的 Grok 模型 API，强调实时知识、幽默和最大限度地追求真相，可访问 X 平台数据以获取时事和无审查响应。

阿里巴巴的综合 AI 平台，提供 Qwen 系列模型，具有强大的多语言和多模态能力，与阿里巴巴云生态系统集成。

使用大规模 Cerebras 晶圆级芯片的超快速推理平台，提供最高吞吐量的 LLM 服务。

Google's Gemini API (via AI Studio and Vertex AI) offers multimodal models with massive context windows, strong reasoning, and seamless integration with Google Cloud services for enterprise-scale AI applications.

Mistral AI offers high-performance open-weight and proprietary models via API, optimized for efficiency, coding, and multilingual tasks with strong European data sovereignty focus.

Cohere provides enterprise-focused LLM APIs for text generation, embeddings, and reranking with strong emphasis on business use cases, safety, and customizable models.

Perplexity AI API provides access to search-enhanced LLMs with real-time web knowledge and citations for accurate, up-to-date responses.

Hugging Face provides Inference Endpoints and API for thousands of open models, with serverless options and enterprise hosting for seamless deployment and fine-tuning.

智谱AI的GLM系列模型以高性价比和强大推理能力著称。

Public AI offers accessible inference endpoints for open-source LLMs with community focus.

01.AI的Yi系列模型在多语言和编码任务中表现突出。

N/A

N/A

Alibaba's Qwen series offers strong multilingual (especially Chinese) and coding capabilities through cloud APIs. Competitive performance with open-weight options available.

01.AI's Yi models are open-source focused with strong bilingual capabilities and competitive performance across tasks.

ShareAI offers multi-provider routing, observability, and bring-your-own-inference for comprehensive LLM management.

Kong AI Gateway unifies access across multiple LLM providers with built-in RAG, security, and PII protection features for production API management.

阿里的 Qwen（通义千问）系列提供高性能开源和 API 模型，在多语言、编码和多模态任务中表现出色，并与阿里云深度集成。

阿里的 Qwen 系列提供开放和专有模型，具有出色的多语言和编码性能。与阿里云集成以实现企业规模。

Unified platform for serving, deploying, and managing LLMs with focus on production-grade inference and MLOps.

Scaleway offers European sovereign cloud LLM API with free tier and support for open models in EU data centers.

Cloudflare provides AI Gateway for unified, secure, and edge-accelerated access to multiple LLM providers with caching and observability.

通义千问

EU-sovereign AI inference platform with OpenAI-compatible API for European data compliance.

Anthropic delivers safe and reliable Claude models via API, excelling in coding, reasoning, and long-context tasks. Available directly and through Bedrock/Vertex with strong safety features.

谷歌提供 Gemini 模型，通过 AI Studio 和 Vertex AI 交付多模态、长上下文 LLM，与 Google Cloud 深度集成，支持搜索 grounding 和企业安全功能。

Nebius offers cloud AI platform with high-performance LLM inference and GPU infrastructure for developers.

Google provides the Gemini multimodal models through APIs, supporting text, image, audio, and code understanding with strong performance and integration options.