Groq free tier: limits we hit and what answered

Groq — docs

Call it

Base URL: https://api.groq.com/openai/v1
API key: your own from Groq

curl https://api.groq.com/openai/v1/chat/completions \
  -H 'Authorization: Bearer YOUR_GROQ_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{"model":"allam-2-7b","messages":[{"role":"user","content":"hello"}]}'

OPENAI_BASE_URL=https://api.groq.com/openai/v1
OPENAI_API_KEY=your-groq-api-key
OPENAI_MODEL=allam-2-7b

Using allam-2-7b, answered all 42 checks we made in the last week, last answered our key 2 h ago from this server's address. Every model this endpoint offers is in the table below.

Track record: Groq

580 of 1008 real chat calls answered in 14 days (58%), median 281 ms.

The 428 that did not answer were 420 refused for the one model asked about (HTTP 4xx: listed by the provider, missing or not on this tier), 8 server errors (5xx).

Counting only the 588 calls the provider accepted for a model it offers us, 580 answered (99%): the rest of the gap is models that are listed and refused, not the provider being down.

Median reply 281 ms, 95th percentile 456 ms, over the calls that answered. We have been calling this endpoint since 4 August 2026. Measured with our own free-tier key from this server's address: it says the endpoint answered our account, not what a new signup receives.

Day (UTC)	Chat calls	Answered	Median reply
26 Oct	72	38 of 72	282 ms
27 Oct	72	40 of 72	281 ms
28 Oct	72	42 of 72	322 ms
29 Oct	72	42 of 72	326 ms
30 Oct	72	42 of 72	269 ms
31 Oct	72	42 of 72	264 ms
1 Nov	72	42 of 72	267 ms
2 Nov	72	42 of 72	293 ms
3 Nov	72	42 of 72	296 ms
4 Nov	72	42 of 72	298 ms
5 Nov	72	42 of 72	232 ms
6 Nov	72	42 of 72	324 ms
7 Nov	72	42 of 72	297 ms
8 Nov	21	12 of 21	329 ms

This provider requires an API key. Every row below was measured with stillworks' own free-tier key, from this server's address, at the time shown. That says the endpoint answered our account — it says nothing about what a new signup's free tier includes, and we have no way to measure that for you. Our key is never shown here, returned by the API, or stored in the database.

Free tier, key required. Account opened 2026-08-04 with Google sign-in only: no card, no phone number, no verification email, no onboarding form. Listing read with our key on 2026-08-04 returned 15 models, including Whisper speech-to-text and Orpheus speech models that no chat probe will ever touch. Published free-plan limits on that date: 30 requests per minute across chat models, and per-day caps from 250 (groq/compound) through 1,000 (llama-3.3-70b-versatile, the gpt-oss pair, qwen3.6-27b) to 14,400 (llama-3.1-8b-instant). Three probes an hour sits far inside all of those.

Models

Model	Tier	Answers our key	Chat	Tools	Schema	JSON	Vision	Draws	Worked	Last checked	Last listed
`allam-2-7b`	—	yes, on our key	yes	no	no	yes	no	—	42/42	2 h ago	25 min ago
`groq/compound`	—	yes, on our key	yes	no	no	no, claimed	no	—	0/42	25 min ago	47 days ago
`groq/compound-mini`	—	yes, on our key	yes	no	no	no, claimed	no	—	0/42	1 h ago	47 days ago
`llama-3.1-8b-instant`	—	yes, on our key	yes	yes	no	yes	no	—	0/42	3 h ago	82 days ago
`llama-3.3-70b-versatile`	—	yes, on our key	yes	yes	no	yes	no	—	0/42	25 min ago	82 days ago
`meta-llama/llama-prompt-guard-2-22m`	—	yes, on our key	yes	no	no	no	no	—	42/42	2 h ago	25 min ago
`meta-llama/llama-prompt-guard-2-86m`	—	yes, on our key	yes	no	no	no, claimed	no	—	42/42	2 h ago	25 min ago
`openai/gpt-oss-120b`	—	yes, on our key	yes	yes	yes	yes	no	—	42/42	1 h ago	25 min ago
`openai/gpt-oss-20b`	—	yes, on our key	yes	yes	yes	yes	no	—	42/42	3 h ago	25 min ago
`openai/gpt-oss-safeguard-20b`	—	yes, on our key	yes	yes	yes	yes	no	—	42/42	25 min ago	25 min ago
`qwen/qwen3.6-27b`	—	yes, on our key	yes	yes	no	no, claimed	yes	—	0/42	1 h ago	54 days ago
`qwen/qwen3.8-27b`	—	yes, on our key	yes	yes	yes	yes	no	—	42/42	3 h ago	25 min ago
`canopylabs/orpheus-arabic-saudi`	—	not yet checked	not a chat model	—	—	—	—	—	—	never	25 min ago
`canopylabs/orpheus-v1-english`	—	not yet checked	not a chat model	—	—	—	—	—	—	never	25 min ago
`whisper-large-v3`	—	not yet checked	not a chat model	—	—	—	—	—	—	never	25 min ago
`whisper-large-v3-turbo`	—	not yet checked	not a chat model	—	—	—	—	—	—	never	25 min ago

Scroll the table sideways for the rest of each row. Chat, Worked and Last listed are folded away at this width.

Feature columns are measurements. Where the provider's own listing claims a feature we could not get out of it, the cell says so — that disagreement is the reason this shelf exists.

What Tools, Schema, JSON and Vision mean, and what each cell state means

The columns

Each name links to the call this endpoint's verdict was read from.

Tools: We sent a tools array and the model came back with a well-formed tool_calls entry an agent could dispatch. A 200 with no tool call in it is not tool support, and is recorded as a no.
Schema: The strict structured-output mode: response_format: {"type": "json_schema", …}. The reply parsed and matched the schema we sent.
JSON: The weaker mode: response_format: {"type": "json_object"}. Valid JSON, with no shape enforced. The two come apart in both directions — OVH's Mistral-7B-Instruct-v0.3 does the strict form and emits invalid JSON for the weak one; llm7's meta-Llama-3.1-8B-Instruct-Turbo does json_object while json_schema answers 405.
Vision: A real image went in inline and a correct answer about it came back. We ask about something only a model that looked could know.
Draws: On an endpoint's own page: bytes came back with an image content type, and they really are an image. Not a 200 taken on trust.

The cell states

yes: A real call proved it, at the time the endpoint page prints beside that call. The dot's tooltip says the same thing where there is a pointer to hover with.
no: A clean negative: the parameter was refused, or the answer did not conform.
no, claimed: The negative contradicts the provider's own published claim — their listing advertises the feature and our call could not get it out of the model. This is the most interesting cell on the site and the reason the shelf exists.
never probed: No verdict. Either nobody has asked yet, or the check ran and got no usable answer back — Pollinations sits here because its anonymous pool answered 402 to every feature probe. It is a question, never a no.

The same columns, in the API

curl 'https://stillworks.supercapybara.com/api/llm/up?feature=tools,json_schema'

The names are the API's spelling of the columns: tools, json_schema, json_object, vision, image_out. Several are and, never or — a row comes back only if a real call proved every feature named. Models we never probed for one of them are left out rather than optimistically included, so a filtered shelf is shorter than a hopeful one.

The call behind each verdict

One row per feature: the most recent call that decided that column, whatever it answered. A refusal is the evidence behind a no exactly as a success is the evidence behind a yes, so both are here.

Feature	When	Outcome	HTTP	Latency	Model	Detail
Tools	13 h ago	client_error	404	121 ms	`groq/compound-mini`	The model `groq/compound-mini` does not exist or you do not have access to it.
Schema	13 h ago	client_error	400	116 ms	`allam-2-7b`	This model does not support response format `json_schema`. See supported models at https:/…
JSON	13 h ago	ok	200	365 ms	`allam-2-7b`
Vision	13 h ago	client_error	400	117 ms	`allam-2-7b`	messages[0].content must be a string

Raw probe log

Our working, so you can check it rather than trust it. Wider than the page at any width — scroll it sideways for the Detail column, which quotes the provider's own refusal back.

When	Kind	Outcome	HTTP	Latency	Model	Detail
25 min ago	chat	client_error	404	116 ms	`groq/compound`	The model `groq/compound` does not exist or you do not have access to it.
25 min ago	chat	ok	200	376 ms	`openai/gpt-oss-safeguard-20b`
25 min ago	chat	client_error	404	114 ms	`llama-3.3-70b-versatile`	The model `llama-3.3-70b-versatile` does not exist or you do not have access to it.
25 min ago	models	ok	200	129 ms	`—`
1 h ago	chat	client_error	404	108 ms	`groq/compound-mini`	The model `groq/compound-mini` does not exist or you do not have access to it.
1 h ago	chat	client_error	404	111 ms	`qwen/qwen3.6-27b`	The model `qwen/qwen3.6-27b` does not exist or you do not have access to it.
1 h ago	chat	ok	200	304 ms	`openai/gpt-oss-120b`
1 h ago	models	ok	200	132 ms	`—`
2 h ago	chat	ok	200	369 ms	`allam-2-7b`
2 h ago	chat	ok	200	151 ms	`meta-llama/llama-prompt-guard-2-86m`
2 h ago	chat	ok	200	148 ms	`meta-llama/llama-prompt-guard-2-22m`
2 h ago	models	ok	200	136 ms	`—`
3 h ago	chat	client_error	404	363 ms	`llama-3.1-8b-instant`	The model `llama-3.1-8b-instant` does not exist or you do not have access to it.
3 h ago	chat	ok	200	458 ms	`openai/gpt-oss-20b`
3 h ago	chat	ok	200	165 ms	`qwen/qwen3.8-27b`
3 h ago	models	ok	200	151 ms	`—`
4 h ago	chat	client_error	404	112 ms	`groq/compound`	The model `groq/compound` does not exist or you do not have access to it.
4 h ago	chat	ok	200	1069 ms	`openai/gpt-oss-safeguard-20b`
4 h ago	chat	client_error	404	112 ms	`llama-3.3-70b-versatile`	The model `llama-3.3-70b-versatile` does not exist or you do not have access to it.
4 h ago	models	ok	200	132 ms	`—`
5 h ago	chat	client_error	404	127 ms	`groq/compound-mini`	The model `groq/compound-mini` does not exist or you do not have access to it.
5 h ago	chat	client_error	404	127 ms	`qwen/qwen3.6-27b`	The model `qwen/qwen3.6-27b` does not exist or you do not have access to it.
5 h ago	chat	ok	200	472 ms	`openai/gpt-oss-120b`
5 h ago	models	ok	200	130 ms	`—`
6 h ago	chat	ok	200	353 ms	`allam-2-7b`
6 h ago	chat	ok	200	168 ms	`meta-llama/llama-prompt-guard-2-86m`
6 h ago	chat	ok	200	156 ms	`meta-llama/llama-prompt-guard-2-22m`
6 h ago	models	ok	200	133 ms	`—`
7 h ago	chat	client_error	404	113 ms	`llama-3.1-8b-instant`	The model `llama-3.1-8b-instant` does not exist or you do not have access to it.
7 h ago	chat	ok	200	315 ms	`openai/gpt-oss-20b`
7 h ago	chat	ok	200	162 ms	`qwen/qwen3.8-27b`
7 h ago	models	ok	200	141 ms	`—`
8 h ago	chat	client_error	404	113 ms	`groq/compound`	The model `groq/compound` does not exist or you do not have access to it.
8 h ago	chat	ok	200	269 ms	`openai/gpt-oss-safeguard-20b`
8 h ago	chat	client_error	404	115 ms	`llama-3.3-70b-versatile`	The model `llama-3.3-70b-versatile` does not exist or you do not have access to it.
8 h ago	models	ok	200	131 ms	`—`
9 h ago	chat	client_error	404	125 ms	`groq/compound-mini`	The model `groq/compound-mini` does not exist or you do not have access to it.
9 h ago	chat	client_error	404	126 ms	`qwen/qwen3.6-27b`	The model `qwen/qwen3.6-27b` does not exist or you do not have access to it.
9 h ago	chat	ok	200	395 ms	`openai/gpt-oss-120b`
9 h ago	models	ok	200	140 ms	`—`