硅基流动 SiliconFlow
AI Chat ModelsSiliconFlow is a one-stop cloud platform for large language models, offering API access to leading open-source models, serverless inference acceleration and managed model deployment. It aggregates hundreds of models including DeepSeek, Qwen, GLM, Kimi and MiniMax behind an OpenAI-compatible API, with pay-as-you-go billing, a free quota, fine-tuning and private deployment. It plugs directly into tools such as Claude Code, Cursor and Dify.
SiliconFlow is a one-stop cloud platform for large language models built for developers and teams: a unified API that aggregates mainstream open-source models such as DeepSeek, Qwen, GLM, Kimi and MiniMax, plus serverless inference acceleration, fine-tuning and private deployment. It is OpenAI-compatible and billed by usage with a free quota, making it the infrastructure layer of choice for building AI applications and comparing models.
- One-stop model API: call hundreds of leading and open-source models such as DeepSeek, Qwen, GLM, Kimi, MiniMax and Llama through a single endpoint
- OpenAI-compatible: the API follows the OpenAI specification, so migrating only requires changing the base URL and key; Claude Code, Cursor, Dify and Cherry Studio connect in one step
- Serverless inference acceleration: an in-house inference engine with dynamic batching keeps latency low under high concurrency, with no need to build your own GPU cluster
- Pay-as-you-go with no barrier: billed by actual token usage, with no minimum spend and no subscription lock-in, plus a free quota for new users
- Fine-tuning and private deployment: LoRA and other fine-tuning methods plus dedicated instances for domain-specific customization and data isolation
- Multimodal coverage: language, speech, vision, image generation, embeddings and reranking models — one platform for many scenarios
- Cost and performance insight: the console breaks down call volume, latency and spend per model for easy cost benchmarking and optimization
- High availability and compliance: multi-zone redundancy, enterprise-grade SLA and data encryption, with direct access from mainland China for long-running production use
商量SenseChatChatGPT-like artificial intelligence large language model launched by SenseTime
JanTool for running large models locally and conducting AI conversations, free and open source (Overseas service)
ChatGPTOpenAI AI conversational assistant, supports natural language interaction, code writing, content creation multi-turn reasoning
ClaudeAn AI assistant developed by Anthropic, good at long document analysis, code writing, creative writing and rigorous reasoning
GeminiThe multi-modal AI assistant launched by Google DeepMind supports text, code, image, and video understanding and generation
Qwen ChatThe free AI dialogue assistant launched by Alibaba Tongyi Qianwen is based on the Qwen series model and supports in-depth thinking, image and text understanding, web development and in-depth research.
PoeThe AI chat platform launched by Quora integrates top models such as ChatGPT/Claude/GPT-4 and supports fast switching.
华为小艺Huawei's Hongmeng system-level AI smart assistant integrates Pangu large model and DeepSeek to support voice dialogue, multi-modal understanding, intelligent agents and cross-device service flow.