Every row below is machine-verified. Legal free tiers, trials and free-model APIs for AI coding — probed twice a week against live model APIs and pricing pages; dead offers drop to the Archive automatically.
| Tool | What you get | Free models | Limits | Card required | Verified |
|---|---|---|---|---|---|
| opencode | Open-source TUI/desktop coding agent with free models included via the opencode Zen gateway (Big Pickle, DeepSeek V4 Flash, MiMo-V2.5, Nemotron 3 Ultra); any provider via BYOK too | big-pickle, deepseek-v4-flash, mimo-v2.5, nemotron-3-ultra | Bundled Zen models priced Free (some marked limited-time); frontier models pay-as-you-go | ✅ No | 2026-07-23 |
| Kilo Code | VS Code agent extension; free starter credits usable on 500+ routed models (Claude Sonnet 5, GPT-5.5, Gemini 3.1 Pro, ...) | — | Free credits on signup | ✅ No | 2026-07-23 |
| MiMo Code | Xiaomi's open-source (MIT) terminal coding agent; built-in "MiMo Auto" channel runs Xiaomi's MiMo models with no login, API key, or card | mimo-v2.5 | MiMo Auto is free for a limited time; anonymous channel, zero configuration | ✅ No | 2026-07-23 |
| OpenAI Codex CLI | Open-source coding CLI, free by signing in with a $0 ChatGPT Free account; local coding tasks included on all plans | gpt-5.6 | Free ChatGPT plan carries the smallest allowance; shared 5-hour rolling + weekly rate limits; local tasks only | ✅ No | 2026-07-23 |
| Crush + Charm Hyper | Charm's Crush terminal agent with Hyper, its official hosted model gateway; the free plan includes monthly Hypercredits, zero data retention | — | 100 Hypercredits (≈$5) refreshed monthly; Hyper is in private beta (sign up from Crush or the site) | ✅ No | 2026-07-23 🧪 |
| Tool | What you get | Free models | Limits | Card required | Verified |
|---|---|---|---|---|---|
| NVIDIA NIM (build.nvidia.com) | Free hosted NIM endpoints for 100+ models via the free NVIDIA Developer Program (OpenAI-compatible at integrate.api.nvidia.com/v1) | nemotron, llama-4 | Free tier ~40 req/min, no credit card; production use needs NVIDIA AI Enterprise | ✅ No | 2026-07-23 |
| Groq | Fast inference free tier | llama-4, qwen3 | Free tier daily limits per model | ✅ No | 2026-07-23 |
| Cerebras Inference | Very fast inference, free tier | qwen3 | Free tier with daily limits | ✅ No | 2026-07-23 |
| OVHcloud AI Endpoints | EU-hosted serverless open-model API; anonymous tier needs no signup or API key (OpenAI-compatible) | qwen3, gpt-oss | No-key anonymous access, rate-limited; free API key raises limits | ✅ No | 2026-07-23 |
| Z.ai (Zhipu GLM) | GLM Flash models free on the API (OpenAI-compatible at api.z.ai/api/paas/v4) | glm-4.5 | GLM-4.7-Flash / GLM-4.5-Flash / GLM-4.6V-Flash priced Free; flagship GLM-5.x not free; rate-limited | ✅ No | 2026-07-23 |
| LongCat API Platform | Meituan's LongCat platform — OpenAI- and Anthropic-compatible API with a recurring daily free token quota | longcat-2.0, longcat-flash | 100K free tokens/day; a larger free tier for Flash-Lite (50M tokens/day) is announced and the model is uncapped during rollout | ✅ No | 2026-07-23 🧪 |
| Cloudflare Workers AI | 10k neurons/day free | llama-4 | 10,000 neurons/day free allocation | ✅ No | 2026-07-23 |
| Ollama Cloud | Cloud-hosted open models with free usage tier | nemotron, minimax-3, gpt-oss | Free tier with hourly/daily limits; open models only — flagship models (DeepSeek V4, GLM-5, Kimi K2.x, Qwen3.5) need a subscription | ✅ No | 2026-07-23 🧪 |
| GitHub Models | Free playground+API for catalog models with GitHub account | gpt-4.1 | Per-model rate limits, free tier | ✅ No | 2026-07-23 |
| Mistral La Plateforme | Free experiment tier on La Plateforme | mistral-medium | Experiment tier rate limits | ✅ No | 2026-07-23 |
| Reka AI | Reka multimodal API (chat, research, vision) with free credits refreshed automatically every month | reka-core, reka-flash | $10 free credits at the start of every month, usable on any API feature | ✅ No | 2026-07-23 🧪 |
| Pollinations.AI | Open GenAI text API, no signup, OpenAI-compatible (POST text.pollinations.ai/openai) | gpt-oss | Anonymous 1 req/15s (no signup); free registration 1 req/5s; anon text model is GPT-OSS-20B | ✅ No | 2026-07-23 |
| Alibaba Cloud Model Studio (DashScope, international) | Free quota for Qwen models on DashScope, international (Singapore) region; OpenAI-compatible | qwen3-max, qwen3-coder | 1,000,000 free tokens per model, valid 90 days after activation; Singapore/international scope only | ✅ No | 2026-07-23 |
| Cohere (trial keys) | Cohere Command models via free trial API keys that never expire | command-a | Trial key: 1,000 API calls/month, rate-limited; production keys unlock paid volume | ✅ No | 2026-07-23 |
| Novita AI | Inference cloud for 200+ open models; selected models priced Free plus a small signup trial credit | — | Selected models free on the pricing page (e.g. Hunyuan 3); ~$0.5 trial credit valid 1 year | ✅ No | 2026-07-23 🧪 |
| Scaleway Generative APIs | EU-made serverless LLM API (OpenAI-compatible); 1M free tokens for every new customer | glm-5.2, qwen3 | 1,000,000 free tokens then pay-per-token; a valid payment method is required | 💳 Yes | 2026-07-23 |
| Google AI Studio (Gemini API) | Free tier for Gemini 2.5 Flash/Pro API | gemini-2.5 | Low per-model daily caps on the free tier (see rate-limits page) — among the stingiest here | ✅ No | 2026-07-23 |
| Tool | What you get | Free models | Limits | Card required | Verified |
|---|---|---|---|---|---|
| GitHub Copilot Free | Free Copilot plan for individual developers in VS Code, JetBrains, Visual Studio and CLI; completions, limited chat and agent usage | — | 2,000 code completions/month; limited chat & agent requests; auto model selection only | ✅ No | 2026-07-23 |
| DeepSeek Platform | Official DeepSeek API; every new account gets a one-time free token grant, no card required | deepseek-chat, deepseek-reasoner | 5M free tokens on signup (~30-day validity), then pay-as-you-go; no perpetually free models | ✅ No | 2026-07-23 🧪 |
| Kiro | Perpetual free tier of AWS's spec-driven agentic IDE (successor to Amazon Q Developer) with Claude Sonnet 4.5 and open-weight models | claude-sonnet-4.5, qwen3-coder | 50 credits/month; requires social login or AWS Builder ID; credits do not roll over | ✅ No | 2026-07-23 |
| Google Jules | Free tier of Google's async cloud coding agent powered by Gemini 2.5 Pro; connects to GitHub repos and works autonomously | gemini-2.5 | 15 tasks per rolling 24 hours; 3 concurrent tasks | ✅ No | 2026-07-23 |
| Cursor (Hobby) | Permanent free Hobby plan of the Cursor AI IDE with limited Agent requests and Tab completions, no credit card | — | Limited Agent requests and Tab completions; Auto model only; pauses at cap until reset | ✅ No | 2026-07-23 |
| Windsurf | Free plan + trial of paid tiers | claude-haiku, gpt-5.2-mini, kimi-k2.5 | Free plan credits | ✅ No | 2026-07-23 |
| Qoder | Alibaba's agentic coding IDE + CLI; Pro trial with credits on signup, then a free plan with basic models | — | Trial: 300 credits for 2 weeks; free plan afterwards serves basic models with a daily cap | ✅ No | 2026-07-23 🧪 |
| Trae | Free access to frontier models in IDE | — | Free tier quotas | ✅ No | 2026-07-23 |
| Upstage (Solar API) | Upstage Solar LLM API; $10 free credit on signup, no card | solar-pro-3, solar-mini | $10 signup credit (see console for validity); pay-as-you-go after | ✅ No | 2026-07-23 🧪 |
| Tool | What you get | Free models | Limits | Card required | Verified |
|---|---|---|---|---|---|
| OpenRouter (free models) | One API key for rotating :free variants of frontier models | gpt-oss, nemotron-3-ultra, gemma-4 | 50 req/day free (1000/day with $10 balance) | ✅ No | 2026-07-23 |
| Hugging Face Inference Providers | Routed access to 200+ models across providers (Groq, Cerebras, Together, etc.) with a free HF account | deepseek, qwen3 | Free users get $0.10/month credits (subject to change); credits apply only on HF-routed requests | ✅ No | 2026-07-23 |
| 🧪 — added recently on fresh evidence; stays provisional until two weeks of probes confirm it. |
Outdated or unverifiable entries: Empty — the radar is clean.
Connection details for every live OpenAI-compatible API above — paste the base URL into opencode, Codex CLI, aider, Cline or any OpenAI SDK:
| Provider | Base URL | Key env var | Get a key |
|---|---|---|---|
| NVIDIA NIM (build.nvidia.com) | https://integrate.api.nvidia.com/v1 |
NVIDIA_NIM_API_KEY |
key |
| OpenRouter (free models) | https://openrouter.ai/api/v1 |
OPENROUTER_API_KEY |
key |
| DeepSeek Platform | https://api.deepseek.com |
DEEPSEEK_API_KEY |
key |
| Groq | https://api.groq.com/openai/v1 |
GROQ_API_KEY |
key |
| Hugging Face Inference Providers | https://router.huggingface.co/v1 |
HUGGINGFACE_INFERENCE_API_KEY |
key |
| Cerebras Inference | https://api.cerebras.ai/v1 |
CEREBRAS_API_KEY |
key |
| OVHcloud AI Endpoints | https://oai.endpoints.kepler.ai.cloud.ovh.net/v1 |
— | not needed |
| Z.ai (Zhipu GLM) | https://api.z.ai/api/paas/v4 |
ZAI_GLM_API_KEY |
key |
| LongCat API Platform | https://api.longcat.chat/openai/v1 |
LONGCAT_API_KEY |
key |
| Cloudflare Workers AI | https://api.cloudflare.com/client/v4/accounts/{account_id}/ai/v1 |
CLOUDFLARE_WORKERS_AI_API_KEY |
key |
| Ollama Cloud | https://ollama.com/v1 |
OLLAMA_CLOUD_API_KEY |
key |
| GitHub Models | https://models.github.ai/inference |
GITHUB_MODELS_API_KEY |
key |
| Upstage (Solar API) | https://api.upstage.ai/v1 |
UPSTAGE_API_KEY |
key |
| Mistral La Plateforme | https://api.mistral.ai/v1 |
MISTRAL_API_KEY |
key |
| Reka AI | https://api.reka.ai/v1 |
REKA_API_KEY |
key |
| Pollinations.AI | https://text.pollinations.ai/openai |
— | key |
| Alibaba Cloud Model Studio (DashScope, international) | https://dashscope-intl.aliyuncs.com/compatible-mode/v1 |
ALIBABA_MODEL_STUDIO_API_KEY |
key |
| Cohere (trial keys) | https://api.cohere.com/compatibility/v1 |
COHERE_API_KEY |
key |
| Novita AI | https://api.novita.ai/openai |
NOVITA_API_KEY |
key |
| Scaleway Generative APIs | https://api.scaleway.ai/v1 |
SCALEWAY_GENERATIVE_API_KEY |
key |
| Google AI Studio (Gemini API) | https://generativelanguage.googleapis.com/v1beta/openai/ |
GOOGLE_AI_STUDIO_API_KEY |
key |
| Zero-signup sanity check — works with no account, no key, right now: |
curl -s https://oai.endpoints.kepler.ai.cloud.ovh.net/v1/chat/completions \
-H 'Content-Type: application/json' \
-d '{"model":"gpt-oss-120b","messages":[{"role":"user","content":"2+2?"}]}'Ready-made artifacts, regenerated on every update:
configs/opencode.json— drop-in opencode config with every provider wired up (keys via{env:...}, keyless endpoints work immediately)configs/free-llm.env.example— commented env exports for any OpenAI-compatible toolindex.json— machine-readable registry;curl -s https://raw.githubusercontent.com/mvalentsev/awesome-free-ai-coding/main/index.json | jq '.entries[].id'
This repository is an autonomous system, not a hand-curated list:
flowchart LR
S["🌐 web scout<br/>Tavily · HN · GitHub · curated feeds"] --> L["🧠 LLM extract<br/>(evidence only)"]
L --> G{"🛡 probe gate"}
G -->|verified| PR["📬 pull request"]
PR -->|human merge| R[("registry.yaml")]
R --> P{"📡 live probes<br/>twice a week"}
P -->|pass| V["✅ verified date updated"]
P -->|fail ×3 or stale 60d| A["🗄 Archive"]
R -->|render| MD["README.md<br/>(this page)"]
- Live probes, twice a week. GitHub Actions hits every entry's public models API or pricing page and re-verifies the free offer.
Verifieddates are earned by passing a probe, never typed by hand. - Web-evidence scout. A discovery layer sweeps Tavily search, Hacker News, GitHub and curated feeds; an LLM extracts candidates strictly from fetched page evidence — it has no authority to invent anything.
- Probe-gated proposals. Every candidate must pass its own live probe before it is even proposed, and lands only through a reviewable pull request. The LLM never writes to this README or to
main. - Self-pruning. Entries that keep failing probes or stay unverified for 60+ days move to the Archive automatically. Rejected-for-cause domains live in
blocklist.yaml. - Zero-secret resilient. The scout's LLM chain falls back across providers down to a keyless anonymous endpoint, so the pipeline keeps running even with no API keys configured.
registry.yaml is the single source of truth; this README is generated from it — don't edit it by hand.
Know a legal free offer that's missing? Suggest a service — it will be probed like everything else. Details in CONTRIBUTING.md.
⭐ If this list saved you a credit-card form, star the repo — it keeps the radar visible.
Maintained by robots · reviewed by humans · MIT