# free.ai

> free.ai | Inference; 1 record: free tier, limits, identity cost and evidence.

HTML https://fshy.ajam.dev/p/free/ | JSON https://fshy.ajam.dev/api/v1/entity/free.json | how to read a record: https://fshy.ajam.dev/llms-full.txt

## free.ai

- `inference/free-ai` | qualifies-with-caveat | evidence measured | binding none | category inference endpoint, keyless
- needs card no, phone no, ID no, institution no | verified 2026-10-28T05:28:44.329Z (live) | stale on 2026-11-28
- url https://api.free.ai/v1
- free tier: 513 model entries in /v1/models, 45 of them marked self_hosted true; every id probed answers from one backend, Qwen/Qwen3-30B-A3B-Instruct | unit: tokens | metered on: request-count | read on: 2026-10-22 | source: [1] | note: the catalogue carries prompt and completion prices of '0' per entry. The count 513 is the length of the models array, measured
- per-object cap: not applicable
- retention: not applicable
- egress: metered: no
- identity: anonymous | detail: no key required.
- interface: rest
- limits:
  - request-rate: free account API key: 60 requests per minute and 1,000 requests per month; no keyless API limit is published [2]
  - volume-quota: anonymous web use: 6,000 tokens per day, reset at midnight UTC; free account: 30,000 tokens per day [3]
- limits note: the served model is the constraint: 513 catalogue names resolve to one backend, Qwen/Qwen3-30B-A3B-Instruct, so the model cannot be chosen and a capacity claim about free.ai is a claim about that one model. What it does not do is serve what was asked for
- binding note: The failure here is substitution and not a ceiling, and substitution is not one of the six constraint kinds, so it is recorded in limits_note rather than forced into a kind that would misdescribe it
- endpoint: https://api.free.ai/v1/chat/completions
- tooling: any OpenAI-compatible client; curl against POST {base}/chat/completions
- composes with: pairs with a batch job that does not care which model runs
- reliability: no substitution was observed inside a single run, but the model served never matches the model requested for any id other than the backend's own name
- notes: The catalogue is 513 names for approximately one backend. Any capacity or quality claim about free.ai is a claim about Qwen3-30B-A3B, and a benchmark keyed on the requested id would be attributed to a model that never ran.
- evidence note: the substitution was measured three times in run 1 and twice in run 2, plus once with a different id. What remains unmeasured is whether any of the 45 self-hosted entries is served as named.
- probe:
  - records[0].probe_id: freeai-qwen3-8b
  - records[0].requested_model: qwen3-8b
  - records[0].http_run1: 200
  - records[0].http_run2: 200
  - records[0].answered_run1: yes
  - records[0].answered_run2: yes
  - records[0].model_echoed: Qwen/Qwen3-30B-A3B-Instruct
  - records[0].echo_match: SUBSTITUTED
  - records[0].latency_ms_run1: 2360
  - records[0].latency_ms_run2: 1517
  - records[0].ttft_ms_run1: 885
  - records[0].completion_tokens_run1: 9
  - records[0].tps_run1: 10.91
  - records[0].tps_run2: 13.87
  - records[0].needle_echoed_run1: yes
  - records[0].body_bytes_run1: 824
  - records[0].body_sha256_16_run1: 9ef4f43fa87fd7a5
  - records[0].reasoning_only: no
  - records[1].probe_id: freeai-qwen-coder
  - records[1].requested_model: qwen-coder
  - records[1].http_run1: 200
  - records[1].http_run2: 200
  - records[1].answered_run1: yes
  - records[1].answered_run2: yes
  - records[1].model_echoed: Qwen/Qwen3-30B-A3B-Instruct
  - records[1].echo_match: SUBSTITUTED
  - records[1].latency_ms_run1: 1370
  - records[1].latency_ms_run2: 1137
  - records[1].ttft_ms_run1: 819
  - records[1].completion_tokens_run1: 9
  - records[1].tps_run1: 10.03
  - records[1].tps_run2: 12.92
  - records[1].needle_echoed_run1: yes
  - records[1].body_bytes_run1: 826
  - records[1].body_sha256_16_run1: b49b6704165db51e
  - records[1].reasoning_only: no
  - records[2].probe_id: freeai-deepseek-r1
  - records[2].requested_model: deepseek-r1
  - records[2].http_run1: 200
  - records[2].http_run2: 200
  - records[2].answered_run1: yes
  - records[2].answered_run2: no
  - records[2].model_echoed: Qwen/Qwen3-30B-A3B-Instruct
  - records[2].echo_match: SUBSTITUTED
  - records[2].latency_ms_run1: 9223
  - records[2].latency_ms_run2: 6547
  - records[2].ttft_ms_run1: 913
  - records[2].completion_tokens_run1: 225
  - records[2].needle_echoed_run1: yes
  - records[2].body_bytes_run1: 1613
  - records[2].body_sha256_16_run1: bf8d7e91e7bfc069
  - records[2].reasoning_only: no
  - records[2].error_run2: empty content; reasoning field populated (901 chars); finish_reason=length
  - runs: probes-w.json; probes-w-run2.json
- evidence:
  - probe:https://api.free.ai/v1/chat/completions 2026-10-22: qwen3-8b, qwen-coder and deepseek-r1 all echoed Qwen/Qwen3-30B-A3B-Instruct in run 1; qwen3-8b and qwen-coder did so again in run 2
  - [1] 2026-10-22: 513 entries with prompt and completion prices of '0'; 45 carry self_hosted true; the first entries name qwen7b as 'Qwen 3.6 35B' and qwen-coder as 'Qwen 3.6 35B'
  - [2] 2026-10-28: the Free plan row of the API request tiers table: 1,000 requests per month and 60 per minute per key | capture: sha256:d18521cbc1e2fca48572add51568e4ac8352cef3abef18a8c2f36d6d72374f60 | quote: | Free | $0 | 1,000 | 60 |
  - [3] 2026-10-28: anonymous users get 6,000 tokens per day; the same page lists 30,000 per day for a free account and a reset at midnight UTC | capture: sha256:586ff28f499feccd507df08cd26ed66c8d899ffe1c1278cf9828d96617dcdba0 | quote: Anonymous users get 6,000 tokens per day.
  - [2] 2026-10-28: the documented API path needs an email-confirmed account and a Bearer key; the keyless call is not documented | capture: sha256:d18521cbc1e2fca48572add51568e4ac8352cef3abef18a8c2f36d6d72374f60 | quote: Sign up free at /signup/ and confirm your email - that unlocks the 30,000-token daily free pool. Open /account/?tab=api and click Generate.
  - probe:https://api.free.ai/v1/chat/completions 2026-10-28: one POST with no Authorization header for qwen7b answered 200 with content NEEDLE-9K2X and echoed model Qwen/Qwen3-30B-A3B-Instruct (not qwen7b); body free_ai_usage reports tokens_used 37, tokens_charged 37, source self_hosted; no rate-limit header was returned (headers: server nginx/1.18.0, content-length 822)
- sources:
  - [1] https://api.free.ai/v1/models
  - [2] https://free.ai/developer/
  - [3] https://free.ai/tokens/