Groq Free Plan: What You Actually Get Before Paying
Groq's free plan costs $0, needs no card, and has no end date. See the exact per-model limits, what 200,000 tokens a day buys, and when to upgrade.
By Galactic News Editorial · Published 2026-11-08T22:24:20+00:00 · Updated 2026-11-08T22:24:20+00:00
Groq Free Plan: What You Actually Get Before Paying
Groq's free plan costs nothing and needs no credit card. It gives you real API access to fast open-weight models, with limits that are generous enough for learning and small projects. This guide explains exactly what you get, what the limits are, and when you should upgrade to the paid Developer plan.
What the Groq free plan includes
Every GroqCloud account starts on the Free plan. You create an API key and start calling models right away. A payment method is needed only when you upgrade to the Developer plan. The free plan has no documented end date. Groq's billing documentation lets you downgrade back to Free at any time, which only makes sense if Free is permanent.
The plan covers the models in Groq's free rate-limit table. That includes OpenAI's gpt-oss-120b and gpt-oss-20b, Qwen 3.8 27B, Whisper speech-to-text models, and Llama Prompt Guard 2 for safety checks. Llama 3.1 8B and Llama 3.3 70B are Enterprise models now, priced through Groq's sales team. They are not on the free plan.
Groq free plan limits, model by model
Groq sets limits per model. They apply at the organization level, not per user. The limits are measured in requests per minute, requests per day, tokens per minute, and tokens per day. You hit whichever threshold you reach first. The figures below come from Groq's official rate limits documentation, read in November 2026.
| Model | Requests / minute | Requests / day | Tokens / minute | Tokens / day |
|---|---|---|---|---|
| gpt-oss-120b | 30 | 1,000 | 8,000 | 200,000 |
| gpt-oss-20b | 30 | 1,000 | 8,000 | 200,000 |
| qwen3.8-27b | 30 | 1,000 | 8,000 | 200,000 |
| gpt-oss-safeguard-20b | 3 | 1,000 | 2,000 | 200,000 |
| whisper-large-v3 / v3-turbo | 20 | 2,000 | 7,200 audio sec / hour | 28,800 audio sec / day |
| llama-prompt-guard-2 | 30 | 14,400 | 15,000 | 500,000 |
When you exceed a limit, the API returns a 429 Too Many Requests status with a retry-after header in seconds. Wait for the window to reset, then retry. The minute and day quotas refill on their own. Nothing on the plan expires.
What 200,000 tokens a day actually means
The daily token cap is the limit you will hit first. At 200,000 tokens a day on gpt-oss-120b, you can run roughly a few hundred short chats or a few dozen long ones. That is plenty for learning the API, building a prototype, or running a small personal tool. It is not enough for a production app with real users. A single user asking long questions can burn the daily budget in an afternoon.
Note that the free plan has no access to Batch processing or Flex processing. Those are Developer plan features. Batch takes 50 percent off every token. Flex gives higher limits on paid accounts. If you need either, you need to upgrade.
How to get started on the free plan
Getting started takes about five minutes. Follow these steps.
- Create a GroqCloud account. Sign up with an email address. No credit card is needed for the Free plan.
- Create an API key. Open your account settings and generate a key. Keep it secret. Never put it in client-side code or a public repository.
- Make your first call. Send a chat completion request to gpt-oss-120b. Groq's API is OpenAI-compatible, so most OpenAI client libraries work with a changed base URL.
- Watch your usage. Check the response headers for your remaining requests and tokens. The x-ratelimit-remaining-requests and x-ratelimit-remaining-tokens headers show your current budget.
- Pick the right model. Use gpt-oss-20b for fast, cheap calls. Use gpt-oss-120b when you need stronger reasoning. Use Whisper for transcription. Qwen 3.8 27B is the only self-serve model that takes images, but it is a preview model and may be discontinued at short notice.
What you pay when you upgrade
The Developer plan is pay-as-you-go with no subscription fee. You add a payment method, such as a credit card, a US bank account, or a SEPA debit account. The upgrade is immediate and there is no upfront charge. Prices per million tokens are $0.15 input and $0.60 output for gpt-oss-120b, and $0.075 input and $0.30 output for gpt-oss-20b. Cached input costs half price on the gpt-oss models. Qwen 3.8 27B costs $0.80 input and $4.00 output per million tokens.
Billing is progressive. Your first invoice fires when your lifetime Developer usage reaches $1. Later invoices fire at $10, $100, $500, and $1,000. Groq does not bill amounts under $0.50. The Developer plan also raises the limits sharply. On gpt-oss-120b, you get 1,000 requests a minute, 500,000 requests a day, and 250,000 tokens a minute.
Bottom line
The Groq free plan is a real free tier, not a short trial. You get fast models, no card, and no end date, with limits that fit learning and prototypes. Upgrade to the Developer plan when you hit the 200,000-token daily cap regularly, when you need Batch or Flex processing, or when you are shipping to real users. Until then, the free plan costs you nothing and teaches you everything you need to know about the platform.