> Groq
Groq runs LPU inference hardware tuned for raw tokens-per-second. The API offers a documented free developer tier — per-model rate limits such as 30 RPM / 1,000 RPD / 8K TPM on gpt-oss-120b — with full x-ratelimit-* header observability, and paid Developer plans that add Batch and Flex processing plus higher ceilings.
// pros
- Purpose-built LPU inference — thousands of tokens/second on supported models, among the fastest open-model inference available.
- Documented free tier with per-model limits (e.g. gpt-oss-120b: 30 RPM, 1,000 RPD, 8K TPM) — no card required to start.
- Full rate-limit observability: x-ratelimit-* headers on responses, retry-after on 429, and a live per-org limits page.
- Cached tokens don't count against rate limits, so prompt caching raises effective throughput.
- The paid Developer plan adds Batch and Flex processing and higher limits without changing the API shape.
// cons
- Narrow catalog: a handful of open-weight models (gpt-oss, Qwen, Llama guard, Whisper, Orpheus) rather than a broad lineup.
- Free ceilings are tight and per-model — 1,000 requests/day on gpt-oss-120b suits demos and light agents, not production traffic.
- Limits apply per organization, not per key: teammates sharing an account share one pool.
- The published table is explicitly a summary with possible exceptions — the exact limits for your org live in the console.
The fastest free tier for latency-sensitive prototypes: run gpt-oss-120b or qwen3.8-27b inside the free limits and keep a second provider for overflow. Before you depend on the free RPM, plan the move to the Developer tier.
written by Mike Fleming, Uprouter editorial · verified against official sources September 16, 2026
How we estimated this: the free allowance is converted to a dollar value using the provider’s own pay-as-you-go rates for the models a typical developer would reach for. When the allowance is metered in requests, tokens or minutes rather than dollars, we assume median usage of the cheapest capable model. Because assumptions are involved, the figure carries the est. flag — treat it as an order of magnitude, not a quote.
note: Community reported approximate monthly value of free tier rate limits for individual developers.
| Plan | Type | Monthly | Input /1M | Output /1M | Note |
|---|---|---|---|---|---|
| Free tier (source-reported) | free | — | — | — | Free tier has no time limit and no credit card: e.g. Llama 3.1 8B ~14,400 req/day, 6,000 tokens/min; Llama 3.3 70B ~1,000 req/day, 12,000 tokens/min; limits are per-organization and can be raised via the Developer plan. |
| Free Tier | free | — | — | — | Rate-limited access to open models like Llama 3.1 and Qwen. |
| On-demand PAYG | payg | — | $0.050 | $0.080 | Representative rate for Llama-3.1-8b-instant; varies by model. |
Pricing normalized from public sources — always verify with the provider.
// model prices
| Model | Input /1M | Output /1M | Context |
|---|---|---|---|
| Qwen family | — | — | — |
| Whisper Large v3 | — | — | — |
| Llama-3.1-8B-Instruct | — | — | 131k |
| Llama 3.3 70B Instruct | — | — | 131k |
$ Does Groq have a free tier?
Yes — Groq offers a free tier that we value at roughly $50.00 in free credits (estimate — derived from rate-limited usage at pay-as-you-go prices). Always confirm current limits on Groq's official pricing page — free tiers change without notice.
$ How much does Groq cost per 1M tokens?
Groq does not publish simple per-1M-token pricing (it may bill per request, per output, or via a subscription plan). Check their official pricing page for exact figures.
$ Is Groq safe to use?
Uprouter rates Groq as low risk. Low risk: first-party inference on Groq's own hardware with published rate-limit policy and standard terms, and no hidden data-use opt-in for API traffic. The commercial risk is the funnel shape — the free tier is narrow and deliberately throttled to push workloads into paid plans, and the model lineup tracks what Groq chooses to host rather than what you need. Risk ratings are editorial, evidence-based, and never influenced by affiliate relationships.
Raw community sentiment, shown unblended. Votes never feed into objective data fields like price, risk, or status — and never into sort order.
No published reviews yet — be the first.
// more_informationprice history · live status · risk & confidence · code snippets
price_history
No recorded changes yet.
live_status
full history →best-effort probes · not a guarantee · last 15h ago
| Checked | Status | Latency |
|---|---|---|
| 15h ago | up | 300ms |
| 2d ago | up | 455ms |
| 3d ago | up | 279ms |
| 5d ago | up | 235ms |
| 6d ago | up | 205ms |
| 8d ago | up | 216ms |
| 9d ago | up | 243ms |
| 11d ago | up | 252ms |
| 12d ago | up | 250ms |
| 13d ago | up | 297ms |
| 14d ago | up | 242ms |
| 15d ago | up | 281ms |
| 16d ago | up | 256ms |
| 16d ago | up | 231ms |
| 17d ago | up | 261ms |
| 18d ago | up | 175ms |
| 19d ago | up | 213ms |
| 19d ago | up | 274ms |
| 20d ago | up | 295ms |
| 21d ago | up | 282ms |
| 22d ago | up | 290ms |
| 23d ago | up | 199ms |
| 23d ago | up | 329ms |
| 24d ago | up | 312ms |
| 25d ago | up | 281ms |
| 26d ago | up | 303ms |
| 27d ago | up | 257ms |
| 27d ago | up | 186ms |
| 28d ago | up | 174ms |
| 29d ago | up | 262ms |
Uprouter runs its own probes; best-effort, not a guarantee. Probe results depend on our network location and cadence — for production, always use the provider’s own status page.
risk_&_confidence
Low risk: first-party inference on Groq's own hardware with published rate-limit policy and standard terms, and no hidden data-use opt-in for API traffic. The commercial risk is the funnel shape — the free tier is narrow and deliberately throttled to push workloads into paid plans, and the model lineup tracks what Groq chooses to host rather than what you need.
Confidence reflects how much verifiable evidence (official docs, probe history, community reports) backs this entry. Risk is our editorial assessment of operator reliability and terms — not financial advice.
code_snippets
# Groq's own API — get a key from their dashboard
curl https://YOUR-PROVIDER-ENDPOINT/chat/completions \
-H "Authorization: Bearer THEIR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "qwen-family",
"messages": [{ "role": "user", "content": "Hello" }]
}'Direct to the provider — Uprouter receives nothing; verify their current pricing before purchasing.
Ranked by neutral similarity signals only — live status, free-tier shape and free-credit proximity. Commercial relationships never influence this list.