Top Free Coding Models — Ranked by Context Window & Rate Limit
| Provider | Model | Context | Max Output | Modality | Rate Limit | |
|---|---|---|---|---|---|---|
| OpenRouter | OpenAI: gpt-oss-120b (free) | 131K | 131K | See provider page | Details | |
| OpenRouter | OpenAI: gpt-oss-20b (free) | 131K | 8K | See provider page | Details | |
| OpenRouter | Qwen: Qwen3 Coder 480B A35B (free) | 262K | 262K | See provider page | Details | |
| Mistral AI | Codestral | 256K | 256K | ~1 RPS, 500K TPM | Details | |
| Cerebras | gpt-oss-120b | 128K | 8K | 30 RPM, 14,400 RPD, 1M TPD | Details | |
| Kilo Code | x-ai/grok-code-fast-1:optimized:free | 131K | 131K | ~200 req/hr | Details | |
| LLM7.io | qwen2.5-coder-32b | 131K | 131K | 30 RPM (120 with token) | Details | |
| OVHcloud AI Endpoints | Qwen3-Coder-30B-A3B-Instruct | 262K | 32K | 2 RPM (anonymous) | Details |
How to Get Started with Free Coding APIs
Click any model above to get the API key and one-click config for Claude Code (cc), Cursor, and other AI coding tools. All models are permanently free — no credit card required.