Free LLM Collection

这是一个关于免费的大模型api的合集，并精选了一部分模型

This is a collection of free LLM apis, and selected some models

我会尽可能更新维护这个项目（目前只有我一个人）

I will keep maintaining and updating this project to the best of my ability

入选原则是：限制请求速率而不是token > 尽可能多的来源 > 尽可能新且好的模型 > 足够用的请求速率

The selection criteria are: limit request rate over token count > more sources > newer and better models > sufficient rate limits

主要是有一定热度的文本模型

Primarily text models that have gained some popularity

目前只接受提供了OpenAI格式的API

At present, only accepted OpenAI-formated API

欢迎大家分享更多api

Welcome to share more apis

Img Table

这个表格是由Deepseek V4 Flash Thinking生成的，由Taple渲染

This table was generated by Deepseek V4 Flash Thinking, Rendered by Taple

Markdown

硅基流动 / SiliconFlow

• API: https://api.siliconflow.cn/v1
• Rate Limits: 1000 RPM (each model)
• Models:
  • deepseek-ai/DeepSeek-R1-0528-Qwen3-8B
  • Qwen/Qwen3-8B
  • Qwen/Qwen3.5-4B
  • THUDM/glm-4-9b-chat
  • THUDM/GLM-4-9B-0414
  • THUDM/GLM-Z1-9B-0414
  • THUDM/GLM-4.1V-9B-Thinking

OpenRouter

• API: https://openrouter.ai/api/v1
• Rate Limits: 20 RPM / 200 RPD (each model)
• Models:
  • qwen/qwen3-coder:free
  • qwen/qwen3-next-80b-a3b-instruct:free
  • openai/gpt-oss-120b:free
  • nvidia/nemotron-3-nano-30b-a3b:free
  • nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free
  • nvidia/nemotron-3-super-120b-a12b:free
  • arcee-ai/trinity-large-preview:free
  • stepfun/step-3.5-flash:free
  • minimax/minimax-m2.5:free
  • poolside/laguna-m.1:free
  • poolside/laguna-xs.2:free
  • google/gemma-4-26b-a4b-it:free
  • google/gemma-4-31b-it:free

书生 / Intern AI

• API: https://chat.intern-ai.org.cn/api/v1
• Rate Limits: 10 RPM
• Tip: 密钥有效期6个月 / The key is valid for 6 months
• Models:
  • intern-latest
  • intern-s1-mini
  • intern-s1
  • intern-s1-pro
  • internvl3.5-latest
  • internvl3.5-241b-a28b

Google Gemini

• API: https://generativelanguage.googleapis.com/v1beta/openai

• Rate Limits: 5 RPM / 20 RPD

• Models:

  • gemini-3-flash-preview

• Rate Limits: 5 RPM / 20 RPD

• Models:

  • gemini-2.5-flash

• Rate Limits: 15 RPM / 500 RPD

• Models:

  • gemini-3.1-flash-lite-preview

• Rate Limits: 10 RPM / 20 RPD

• Models:

  • gemini-2.5-flash-lite

• Rate Limits: 15 RPM / 1500 RPD

• Models:

  • gemma-4-26b-a4b-it
  • gemma-4-31b-it

Cohere

• API: https://api.cohere.ai/compatibility/v1
• Rate Limits: 20 RPM
• Tip:
  • 绑定支付方式可以使用速率限制更宽松的 Production Key / Binding payment methods can use rate limiting and relaxed Production Key
• Models:
  • command-a-reasoning-08-2025
  • command-a-vision-07-2025

Bigmodel

• API: https://open.bigmodel.cn/api/paas/v4/
• Rate Limits: 只有并发数限制（均为30） / Only the number of concurrent transactions is limited (both 30).
• Models:
  • GLM-4-Flash-250414
  • GLM-4V-Flash
  • GLM-4.1V-Thinking-Flash
  • GLM-4.6V-Flash
  • GLM-4.7-Flash

Github Models

• API: https://models.github.ai/inference
• Rate Limits: 15 RPM / 150 RPD
• Tip:
  • 如果使用 Azure API，可以使用更多模型 / If used Azure API, more models available
• Models:
  • openai/gpt-4.1-nano
  • openai/gpt-4.1-mini
  • openai/gpt-4.1
  • openai/gpt-4o
  • openai/gpt-4o-mini
  • openai/gpt-5-nano
  • openai/gpt-5-mini
  • openai/gpt-5-chat
  • openai/gpt-5

熊猫API

• API: https://api520.pro/v1
• Rate Limits: Unknown
• Tip:
  • 赠送¥100额度 / Gift ¥100 Credit
• Models:
  • 太多了自己看

NVIDIA NIM

• API: https://integrate.api.nvidia.com/v1
• Rate Limits: 40 RPM
• Models:
  • deepseek-ai/deepseek-v3.2
  • deepseek-ai/deepseek-v3.1-terminus
  • z-ai/glm4.7
  • moonshotai/kimi-k2-thinking
  • moonshotai/kimi-k2-instruct-0905
  • qwen/qwen3-coder-480b-a35b-instruct
  • qwen/qwen3.5-122b-a10b
  • stepfun-ai/step-3.5-flash

LLM7

• API: https://api.llm7.io/v1
• Rate Limits: 2 RPS / 20 RPM / 100RPH
• Models:
  • gpt-oss-20b
  • GLM-4.6V-Flash

ModelScope

• API: https://api-inference.modelscope.cn/v1/
• Rate Limits: Unknown
• Tip:
  • 每天2000次 / 2000 times per day
• Models:
  • inclusionAI/Ling-2.6-1T
  • deepseek-ai/DeepSeek-V4-Pro
  • deepseek-ai/DeepSeek-V4-Flash
  • deepseek-ai/DeepSeek-V3.2
  • ZhipuAI/GLM-5.1
  • ZhipuAI/GLM-5
  • MiniMax/MiniMax-M2.5
  • Qwen/Qwen3-235B-A22B-Instruct-2507
  • Qwen/Qwen3-Coder-480B-A35B-Instruct

Kilo Gateway

• API: https://api.kilo.ai/api/gateway
• Rate Limits: 200RPH (Hour)
• Models:
  • google/gemma-4-26b-a4b-it:free
  • inclusionai/ling-2.6-1t:free
  • inclusionai/ling-2.6-flash:free
  • nvidia/nemotron-3-super-120b-a12b:free
  • tencent/hy3-preview:free
  • openrouter/free

HuggingFace

• API: https://router.huggingface.co/v1
• Rate Limits: 300 RPH (Hour)
• Models:
  • deepseek-ai/DeepSeek-V4-Pro:fastest
  • moonshotai/Kimi-K2.6:fastest
  • google/gemma-4-31B-it:fastest
  • zai-org/GLM-5.1:fastest
  • inclusionAI/Ling-2.6-1T:fastest
  • MiniMaxAI/MiniMax-M2.7:fastest
  • deepseek-ai/DeepSeek-V3.2:fastest
  • zai-org/GLM-5:fastest
  • nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16:fastest

Groq

• API: https://api.groq.com/openai/v1/
• Rate Limits: 30 RPM / 1000 RPD
• Models:
  • openai/gpt-oss-120b
  • openai/gpt-oss-20b
  • qwen/qwen3-32b

Celebras

• API: https://api.celebras.ai/v1

• Rate Limits: 30 RPM / 900 RPH / 1440 RPD

• Models:

  • gpt-oss-120b
  • qwen-3-235b-a22b-instruct-2507

• Rate Limits: 10 RPM / 100RPH / 100 RPD

• Models:

  • zai-glm-4.7

Mistral

• API: https://api.mistral.ai/v1
• Rate Limits: Unknown
• Models:
  • mistral-large-2512
  • mistral-small-2603
  • mistral-medium-3.5

LongCat

• API: https://api.longcat.chat/openai/v1
• Rate Limits: Unknown
• Tip:
  • 500,000Tokens Per Day: LongCat-Flash-Chat LongCat-Flash-Thinking LongCat-Flash-Thinking-2601 LongCat-Flash-Omni-2603 LongCat-Flash-Chat-2602-Exp
  • 50,000,000 Tokens Per Day: LongCat-Flash-Lite
  • 5,000,000 Tokens Per Day: LongCat-2.0-Preview
• Models:
  • LongCat-Flash-Chat
  • LongCat-Flash-Thinking
  • LongCat-Flash-Thinking-2601
  • LongCat-Flash-Lite
  • LongCat-Flash-Omni-2603
  • LongCat-Flash-Chat-2602-Exp
  • LongCat-2.0-Preview

OpenCode Zen

• API: https://opencode.ai/zen/v1
• Rate Limit: Unknown
• Models:
  • minimax-m2.5-free
  • nemotron-3-super-free

其它 / Other

来源 / Source

• 自行收集 / Self-collected

• 投稿（B站等） / Contributed by others

Star History

排行榜 / Leaderboards

• llm_benchmark：个人评测榜单，可信度高，而且收录更全 / A personal review list, it is highly credible, and it is more comprehensive
• Artifical Analysis
• LMArena