Home / Inference
g4f.space public provider routes (GPT4Free hosted)
g4f.space/apiInference gateway, anonymous, proof of work creditsinference/g4f-space-public-routesJSONcomparereport an errorprovider JSON
- Free allowance
- Five routes (/api/groq, /api/ollama, /api/pollinations, /api/nvidia, /api/gemini) are listed as none required for an API key, but a chat call answered 402 insufficient_credits asking for proof-of-work credits to earn anonymous usage
- Verdict
- fails: quotathe free allowance is too small to matter
- Cost
- free tier
- Identity
- anonymousno card, no phone
- Binding limit
- volume-quotaa credit balance earned by proof-of-work: an anonymous chat call with none answers 402 insufficient_credits; the credits per call are unpublished
- Evidence
- reported3 sources; a third party said so
- Freshness
- stale on 2026-11-13checked 2026-10-29 (live); stale on 2026-11-13
| Free tier | Five routes (/api/groq, /api/ollama, /api/pollinations, /api/nvidia, /api/gemini) are listed as none required for an API key, but a chat call answered 402 insufficient_credits asking for proof-of-work credits to earn anonymous usage unit mixed; metered on unstated; the credit amount per call is not published; /v1 on the same host is listed as requiring a key from g4f.dev/api_key.html, though GET /v1/models answers 200 with no key |
|---|
| Price | none stated |
|---|
| Limits | - volume-quota (binding) a credit balance earned by proof-of-work: an anonymous chat call with none answers 402 insufficient_credits; the credits per call are unpublished source
|
|---|
| Binding | volume-quota the proof-of-work credit balance gates every anonymous completion |
|---|
| Identity | anonymous no key for the /api/<provider> routes per the g4f.dev table (none required); the price is proof-of-work credits rather than an identity |
|---|
| Egress | not published |
|---|
| Interface | rest |
|---|
| Tooling | any OpenAI-compatible client with base URL https://g4f.space/api/<provider>; the doc says to pass a placeholder key for SDKs that require one |
|---|
| Composes with | pairs with nothing until the credit scheme is understood: an anonymous call with no credit is refused |
|---|
| Reliability | GET /api/pollinations/models returns entries whose message is Error 404 DEPLOYMENT_NOT_FOUND on Vercel, and /api/auto/models returns 404; the groq, ollama, nvidia and gemini listings answer 200 |
|---|
| Notes | the five routes front other providers (groq.com, ollama.com, pollinations.ai, nvidia.com, a hosted Gemini provider) and the listings carry per-model request counters; the routes differ in availability |
|---|
| Evidence note | reported the listings were read by GET; the 402 on a chat call comes from a third-party hourly prober that sends chat calls with no Authorization header. No completion was sent, because probe issues GET and HEAD only |
|---|
Evidence 3
Unknown values 3
$.egress.metered egress metering was not published$.per_object_cap.stated the field does not apply to this kind of offer$.retention.stated the field does not apply to this kind of offer