Providers directory
Explore the registry one provider ecosystem at a time.
Provider pages now have their own crawlable URLs, so the catalog can surface structured coverage by vendor instead of hiding provider discovery behind client-side state.
Sorted by model coverage so the largest provider surfaces stay easy to scan first.
OpenRouter
OpenRouter provides an OpenAI-compatible API gateway for models from multiple vendors, with unified routing, pricing, and availability across providers.
OpenAI
Structured provider summary with direct access to its indexed model catalog.
SiliconFlow
SiliconFlow (硅基流动) is a leading AI infrastructure platform in China, providing optimized inference for 100+ open-source models including DeepSeek, Qwen, GLM, and more. Offers competitive pricing, high performance, and comprehensive model support including LLMs, vision models, image/video generation, embeddings, and audio models.
Infini-AI
Infini-AI (无问芯穹) provides AI computing optimization and large model deployment on diverse chips. They offer GenStudio API service with multiple pre-built LLM models from various vendors.
Qwen
Qwen (通义千问) is the large language model and multimodal model series from the Qwen Team, Alibaba Group.
Cohere
Cohere leads in providing enterprise AI solutions, specializing in Retrieval-Augmented Generation (RAG) and complex tool use. Their Command R series is optimized for long-context industrial applications.
Google's Gemini is a family of multimodal large language models developed by Google DeepMind. Gemini is Google's most capable and general model, built to be natively multimodal and capable of reasoning across text, images, video, and code.
ByteDance
ByteDance's Doubao (豆包) models, available via Volcengine, are designed for enterprise-grade performance with industry-leading cost-effectiveness. The Doubao family includes Pro (flagship), Lite (speed), and specialized models for character, voice, and vision, serving millions of users daily.
Anthropic
Structured provider summary with direct access to its indexed model catalog.
Mistral AI
Mistral AI is a Paris-based frontier AI company founded in 2023, pioneering open-source large language models. Known for their high-performance Mixture of Experts (MoE) architecture, Mistral delivers cutting-edge AI capabilities through models like Mistral Large and Mixtral, combining efficiency with state-of-the-art performance.
Meta
Meta's Llama series represents the standard for high-performance open-weights large language models. With versions ranging from 8B to 405B parameters, Llama enables developers and researchers to build upon frontier-level AI capabilities.
360AI
360AI (360智脑) is an AI platform by Qihoo 360, offering secure and reliable large language models with strong capabilities in Chinese language understanding, reasoning, and content generation.
DeepSeek
DeepSeek AI, founded in 2023 by Liang Wenfeng and backed by the hedge fund High-Flyer, is a Hangzhou-based AI company specializing in advanced coding models. DeepSeek is renowned for delivering exceptional programming capabilities with industry-leading cost-effectiveness through models like DeepSeek-V2.5 and DeepSeek Coder.
MiniMax
MiniMax, founded in 2021, is a Beijing-based AI company specializing in audio and multimodal AI technologies. Known for their text-to-speech and voice cloning capabilities, MiniMax offers models like abab6.5 series that excel in Chinese language understanding and audio generation.
01.AI
01.AI, founded by Kai-Fu Lee, is a global AI company committed to building AI 2.0 systems. Their flagship Yi series models, known for strong bilingual (English/Chinese) performance and long-context capabilities, have achieved top rankings on global evaluation leaderboards.
Moonshot AI
Moonshot AI (月之暗面), founded in March 2023 by Yang Zhilin, Zhou Xinyu, and Wu Yuxin, is a Beijing-based AI company specializing in long-context understanding and logical reasoning. Their flagship Kimi chatbot launched in October 2023, and their models excel at processing extensive documents and complex reasoning tasks.
xAI
xAI, founded by Elon Musk, is dedicated to understanding the true nature of the universe. Their Grok models feature real-time access to information on X and excel in reasoning, math, and coding tasks.
Zhipu AI
Zhipu AI (智谱AI), also known as Z.ai internationally, is a leading Chinese AI company with strong ties to Tsinghua University. The company excels in multi-modal AI capabilities, offering advanced models like GLM-4 for language tasks and CogVLM for vision understanding, positioning itself as China's premier full-stack AI development platform.
Baichuan
Baichuan Intelligence, founded by Wang Xiaochuan, is dedicated to building the best Chinese large language models. The Baichuan series is highly regarded in the open-source community for its performance on Chinese benchmarks and efficient architecture.
Baidu
Baidu's ERNIE (Wenxin) series is one of the earliest and most capable Chinese LLMs. ERNIE 4.0 rivals GPT-4 in Chinese tasks, leveraging Baidu's massive knowledge graph to provide accurate and knowledge-rich responses.
iFlytek
iFlytek, a pioneer in intelligent speech and artificial intelligence, presents Spark Desk (讯飞星火). Leveraging iFlytek's dominance in voice technology, Spark Desk excels in voice interaction, translation, and education-focused language tasks.
SenseTime
SenseTime, a global leader in AI software, offers the SenseNova (日日新) large model series. Known for its roots in computer vision, SenseTime delivers exceptional multimodal capabilities, including advanced image understanding and video analysis integrated with LLMs.
Tencent
Tencent's Hunyuan (混元) is a full-stack large model designed for industrial applications. Integrated widely across Tencent's ecosystem (WeChat, Tencent Cloud), it offers robust capabilities in content generation, logical reasoning, and task execution.
Ant Group
Ant Group's Bailing (百灵) is a multimodal flagship model powering Alipay's intelligent assistant. The open-source Ling series features advanced MoE architecture with Ling-Plus (290B total, 28.8B activated) and Ling-Lite (16.8B total, 2.75B activated) models.