The short answer
- No free tokens: OpenAI bills API usage from prepaid credit. The OpenAI help center says that you need a positive account balance to use OpenAI models.
- A tier named Free: the OpenAI rate limit guide lists a usage tier named Free, for users in an allowed geography. It sets a monthly usage limit of $100. The guide does not describe it as free usage.
- ChatGPT is separate: a free ChatGPT account does not include API access.
- GitHub Models is gone: it gave free rate-limited GPT access with a GitHub token. GitHub retired it on 30 July 2026. Read GitHub Models free API for details.
Legitimate options at no cost
| Option | What you get | Limits |
|---|---|---|
| GPT-OSS on free provider tiers | OpenAI open-weight models (GPT-OSS 120B and 20B) on Groq, Cloudflare, NVIDIA NIM, OVH and others | Rate limits per provider. No OpenAI account needed |
| OpenAI data sharing program | Complimentary daily tokens on traffic that you share with OpenAI | Eligible organizations only. Needs a positive balance. Amounts depend on your usage tier |
| API credits with paid ChatGPT plans | OpenAI has a help article about an API credits benefit for eligible ChatGPT subscriptions | We could not read the full terms. Check the OpenAI help center |
| GPT routes on third-party gateways | Small shared credit on some gateways that resell GPT models | Shared with other users. Can end at any time |
The data sharing program is opt-in. Only the account owner can turn it on. When it is on, OpenAI can use the shared traffic to improve its models. Do not turn it on for private or customer data.
GPT-OSS: the free GPT models
GPT-OSS is a family of open-weight models from OpenAI. Because the weights are open, many providers serve them. Several of those providers have recurring free tiers. This is the most dependable free route to an OpenAI model today.
| Model | Free on | Context | Free limits (best offer) | Tools | Vision |
|---|---|---|---|---|---|
| GPT-OSS 120B | Cloudflare Workers AI, ElectronHub, Groq +2 more | 131K | 30 req/min 1,000 req/day 8,000 tokens/min 200,000 tokens/day ~6M | Yes | No |
| GPT-OSS 20B | Api.Airforce, Cloudflare Workers AI, Groq +5 more | 131K | 30 req/min 1,000 req/day 8,000 tokens/min 200,000 tokens/day ~6M | Yes | No |
| GPT-OSS Safeguard 20B | Groq, LLMTR | 131K | 30 req/min 1,000 req/day 8,000 tokens/min 200,000 tokens/day ~6M | Yes | No |
| GPT-OSS 20B (via Groq) | DreamPrompting | 131K | 100 req/min 5,000 req/day 500,000 tokens/day 500K tokens/day · 5,000 req/day | No | No |
| OpenAI: GPT OSS 20B | ElectronHub | 128K | $0.25/week shared free credits | No | No |
| GPT-OSS 120B (via Groq) | DreamPrompting | 131K | 100 req/min 5,000 req/day 500,000 tokens/day 500K tokens/day · 5,000 req/day | No | No |
Live data from the FreeLLMAPI catalog (version 2026.11.19, checked 19 November 2026). Free limits change often. The catalog is checked every day.
GPT-OSS supports tool calling and a context of about 128K tokens on most providers. It is a good default for agents and coding tools. Read the free GPT-OSS API page for provider details.
Other GPT models in the free catalog
Some third-party gateways give a small recurring credit. You can spend it on closed GPT models, such as GPT mini and nano models. The table shows all GPT rows in today's free catalog.
| Model | Free on | Context | Free limits (best offer) | Tools | Vision |
|---|---|---|---|---|---|
| GPT-5.6 Luna | ElectronHub | 1M | $0.25/week shared recurring credit | Yes | Yes |
| GPT-OSS 20B | Api.Airforce, CLōD, Cloudflare Workers AI +6 more | 131K | 30 req/min 1,000 req/day 8,000 tokens/min 200,000 tokens/day ~6M | Yes | No |
| GPT-OSS 120B | CLōD, Cloudflare Workers AI, ElectronHub +3 more | 131K | 30 req/min 1,000 req/day 8,000 tokens/min 200,000 tokens/day ~6M | Yes | No |
| GPT-OSS Safeguard 20B | Groq, LLMTR | 131K | 30 req/min 1,000 req/day 8,000 tokens/min 200,000 tokens/day ~6M | Yes | No |
| OpenAI: GPT-6 Luna | ElectronHub | 1.1M | $0.25/week shared free credits | No | No |
| OpenAI: GPT OSS 20B | ElectronHub | 128K | $0.25/week shared free credits | No | No |
| Gpt 6 Luna | APINex | 1M | 5 req/min 10M weighted tokens/day shared across APINex free routes; 3x token multiplier; 5 requests/minute/account | No | No |
| GPT 5.4 Mini | ElectronHub | 400K | $0.25/week shared recurring credit | Yes | Yes |
| GPT 5.4 Nano | ElectronHub | 400K | $0.25/week shared recurring credit | Yes | Yes |
| GPT 5 Mini | ElectronHub | 400K | $0.25/week shared recurring credit | Yes | Yes |
| GPT 5 Nano | ElectronHub | 400K | $0.25/week shared recurring credit | Yes | Yes |
| GPT 4.1 | ElectronHub | 1M | $0.25/week shared recurring credit | Yes | Yes |
Showing 12 of 19 matching free models. See all models.
Live data from the FreeLLMAPI catalog (version 2026.11.19, checked 19 November 2026). Free limits change often. The catalog is checked every day.
Read the Free limits column with care. A label such as a weekly shared credit means that many users spend the same small amount. These routes are good for a test of a closed GPT model. They are not good for steady traffic.
Which free GPT option to choose
| Your need | Best free option | Why |
|---|---|---|
| An app or agent that runs every day | GPT-OSS through a pool of free tiers | Several providers serve it, so one limit does not stop you |
| Fast answers for a chat interface | GPT-OSS 20B on a fast provider | Small model, high output speed on free tiers |
| A test of a closed GPT model | A gateway with shared credit | Enough for a few requests, not for steady use |
| Production use of closed GPT models | Paid OpenAI credit | Free routes have no service level and can end |
| Code that uses the Responses API | FreeLLMAPI on /v1/responses | Codex CLI and similar tools need this route |
Free tiers change often. Providers add models, remove models and change limits without notice. The catalog on this site is checked every day, so the tables above show the current free routes.
Use the OpenAI SDK with free models
FreeLLMAPI is an OpenAI-compatible gateway that you run on your own computer. It pools free tiers behind one key. Your OpenAI SDK code needs three changes: the base URL, the key and the model name.
OpenAI SDK pointed at FreeLLMAPI
from openai import OpenAI
client = OpenAI(
base_url="http://localhost:3001/v1",
api_key="<unified-key>",
)
resp = client.chat.completions.create(
model="gpt-oss-120b", # or "auto" to let the router choose
messages=[{"role": "user", "content": "Write a haiku about rate limits."}],
)
print(resp.choices[0].message.content)The gateway also serves /v1/responses, /v1/embeddings and other OpenAI-style routes. When one provider returns 429, the router sends the request to the next provider. The X-Routed-Via header shows which provider answered.
Frequently asked questions
Is the OpenAI API free?
No. OpenAI bills API usage from prepaid credit, and you need a positive balance to use its models. The usage tier named Free is a limit level, not free usage.
How do I get a free OpenAI API key?
You can create a key in the OpenAI platform, but requests need credit. For no-cost GPT access, use GPT-OSS on free provider tiers, which need no OpenAI key.
Is GPT-OSS really free?
The model weights are open. Several providers, for example Groq, Cloudflare Workers AI and OVH AI Endpoints, serve GPT-OSS on recurring free tiers with rate limits.
Can I still use GPT models free on GitHub Models?
No. GitHub retired GitHub Models on 30 July 2026. The inference API is not available to any customer.
Do paid ChatGPT plans include API credits?
A ChatGPT subscription is billed separately from the API. OpenAI describes an API credits benefit for some eligible paid plans. Check the OpenAI help center for the current list of plans.
Related guides
- GitHub Models free API: what happened and what to use now
- Is there a free Claude API?
- Free OpenAI-compatible API base URLs
- Use Codex CLI with free models
- Do free LLM APIs train on your data?
Sources
- OpenAI API docs: Rate limits and usage tiers (checked 20 November 2026)
- OpenAI Help Center: Sharing feedback, evaluation and API data with OpenAI (checked 20 November 2026)
- OpenAI Help Center: API credits benefit for eligible ChatGPT subscriptions (checked 20 November 2026)
- GitHub changelog: GitHub Models is now retired (checked 20 November 2026)
- FreeLLMAPI docs: REST API (checked 20 November 2026)