We use cookies to keep the site working and, with your permission, to measure traffic and show relevant ads. Affiliate links set tracking cookies on the retailer's site when you click them. Cookie policy · Your privacy choices
Best Groq offer right now FREE TIER200,000 free tokens per day on gpt-oss-120bStart FreeSee all 4 offers
Checked on the official page
Billing explained in plain English
You sign up on Groq, not here
Nothing resold, ever
Groq free credits, trials and programs for November 2026
We may earn a commission when you sign up through links here. It never changes what you pay. Details
Groq does not issue public promo codes. Groq doesn’t issue public promo codes. Promotional credits are offered occasionally at hackathons and special events, and those are applied by Groq staff rather than redeemed with a code. Everyone else gets the Free plan limits and the structural discounts: 50% off batch and 50% off cached input. That’s the whole list.
200K tokens and 1,000 requests per dayVerified Nov 2Sample
See details
Who qualifies: Any GroqCloud account on the Free plan, no card required
Free plan accounts get 30 requests per minute, 8,000 tokens per minute, 1,000 requests per day and 200,000 tokens per day on gpt-oss-120b, gpt-oss-20b and qwen3.8-27b, plus 7,200 audio seconds per hour on Whisper. Create a key on the API Keys page and start.
Limits are per organization and per model, reset daily and may differ from the public table. Community support only.
Half of synchronous per-token pricingVerified Nov 2Sample
See details
Who qualifies: Developer plan accounts with a payment method on file
Submit a JSONL file of up to 50,000 requests and choose a completion window between 24 hours and 7 days. Batch pricing is 50% lower than synchronous calls, only completed requests are billed, and batch limits sit outside your per-model rate limits.
Does not stack with the prompt caching discount. Input files up to 200 MB. Results stored for 30 days.
Half price on cached input, no cache feeVerified Nov 2Sample
See details
Who qualifies: All plans, automatic on gpt-oss-20b, gpt-oss-120b and gpt-oss-safeguard-20b
Prompt caching is automatic when a repeated prefix clears the model threshold of 128 to 1,024 tokens. Cached input tokens are billed at 50% and don’t count toward rate limits, which stretches the Free plan’s 8,000 tokens per minute as well as the Developer bill.
Place static instructions first in the prompt. Not combined with the batch discount.
Roughly 10x on-demand rate limits, no surchargeVerified Nov 2Sample
See details
Who qualifies: Paid Developer or Enterprise accounts
Set service_tier to flex in the request body to get about 10x the on-demand rate limits at on-demand prices. Requests fail fast with a 498 capacity_exceeded error when capacity is short, so add jittered retries.
Best effort tier. Not available on the Free plan.
Last checked November 2, 2026. 0 codes, 4 credits, programs and tips live.
Official Groq programs
Every way Groq itself hands out credits or free usage, checked on its own pages. The offers above are the ones worth acting on today.
Anyone who creates a GroqCloud console account. No payment method required. Limits are per organization and per model, and reset daily. Community support only.
Developer plan accounts with a payment method on file. Jobs run within a 24 hour to 7 day window you choose, and only completed requests are billed. The discount does not stack with prompt caching.
Automatic on GPT-OSS 20B, GPT-OSS 120B and GPT-OSS-Safeguard 20B for any plan. Prefix must repeat and clear a model-specific threshold of 128 to 1,024 tokens. Cached tokens do not count toward rate limits.
Roughly 10x on-demand rate limits at the same price
Paid accounts only. Add a card, US bank account or SEPA debit under Settings, Billing. No charge at upgrade; usage is invoiced in arrears at $1, $10, $100, $500 and $1,000 lifetime thresholds, then monthly.
Sign up
Free tier that matters
200,000 tokens per day on gpt-oss-120b, gpt-oss-20b and qwen3.8-27b
8,000 tokens per minute and 30 requests per minute on those models
1,000 requests per day per text model
7,200 audio seconds per hour on whisper-large-v3 and v3-turbo
28,800 audio seconds per day and 2,000 transcription requests per day
100 requests per day on the Orpheus text to speech models
No payment method required, limits reset daily
Recent program changes
MiniMax M2.5 and Qwen3-VL 32B restricted to Enterprise Groq’s changelog says MiniMax M2.5 and Qwen3-VL 32B are available only to Enterprise customers, who must contact their Groq account team for access. Free and Developer accounts can’t call them. Source
GPT-OSS-Safeguard 20B launches with prompt caching Groq’s changelog records the policy-following gpt-oss-safeguard-20b model going live with automatic prompt caching that halves the price of cached input tokens, matching the discount on the other GPT-OSS models. Source
Prompt caching reaches GPT-OSS 120B with limit exemption The changelog entry says automatic prompt caching is live for GPT-OSS 120B, halving cached input costs and excluding cached tokens from rate limits, which effectively raises the Free plan tokens per minute ceiling. Source
Kimi K2-0905 priced at $1.00 input and $3.00 output Groq’s changelog lists Kimi K2-0905 at $1.00 per million input tokens and $3.00 per million output tokens, with prompt caching saving up to 50% on cached portions of the prompt. Source
Groq has no signup credit, no trial dollars and no promo codes. What it has is a Free plan that needs no card and a published rate limit table. On the Free plan gpt-oss-120b, gpt-oss-20b and qwen3.8-27b each get 30 requests per minute, 8,000 tokens per minute, 1,000 requests per day and 200,000 tokens per day. Whisper gets 7,200 audio seconds per hour. That’s enough to build a prototype, not to run one.
Spending on Groq means moving to the Developer plan: pay per token with higher limits, chat support, Flex processing, batch processing and spend limits. The upgrade itself costs nothing. Groq bills in arrears when lifetime usage crosses $1, $10, $100, $500 and $1,000, then monthly. On the Developer plan gpt-oss-120b jumps to 1,000 requests per minute and 250,000 tokens per minute with no daily token cap.
What runs out first on the Free plan is the 8,000 token per minute ceiling, not the daily budget. A single request with a 6,000 token system prompt nearly fills a minute. Prompt caching helps because cached tokens are excluded from rate limits. And the Llama 3.3 70B and 3.1 8B models now show contact sales instead of a price, so plan around the GPT-OSS models.
How to claim Groq credits
Create a GroqCloud accountOpen console.groq.com and sign up with Google, GitHub or email. The account starts on the Free plan with no payment method, and the plans page under Settings, Billing shows Free, Developer and Enterprise side by side.
Create an API keyClick API Keys in the top navigation (the /keys page), create a key and copy it once. Store it as GROQ_API_KEY in your shell rather than pasting it into code, because you won’t see it again.
Test in the PlaygroundOpen Playground from the top navigation, pick openai/gpt-oss-120b, and send a prompt. Playground calls count against the same Free plan limits as API calls, so keep experiments short while you measure tokens.
Read your exact limitsGo to the limits page under Settings to see the rate limits for your organization per model. Where input and output limits differ, hover the TPM value to see the in and out split. These numbers can differ from the public table, and yours win.
Upgrade to Developer when you hit 429sUnder Settings, Billing choose Manage Billing and add a card, US bank account or SEPA debit. The upgrade is immediate with no charge. Then open Limits, click Add Limit, set a monthly USD cap, and add Usage Alerts at 50%, 75% and 90%.
Ways to stretch your Groq credits
Stay on GPT-OSS for price
gpt-oss-20b costs $0.075 input and $0.30 output per million tokens and gpt-oss-120b costs $0.15 and $0.60. Qwen3.8-27b costs $0.80 and $4.00, more than five times gpt-oss-120b on input. Unless you need Qwen specifically, route everyday chat and extraction to the 20B or 120B model and keep Qwen for the calls that earn it.
Batch anything that can wait a day
Batch processing is priced 50% below synchronous calls, with a completion window from 24 hours to 7 days and separate rate limits. Input files take up to 50,000 lines and 200 MB. Evaluations, embeddings backfills and nightly classification belong here. Pick the 7 day window under load to avoid expirations, and half your bulk spend goes away.
Put the static prompt first
Prompt caching halves the price of cached input tokens on GPT-OSS models and costs nothing extra. It only fires when the repeated prefix clears a 128 to 1,024 token threshold, so place system instructions and few-shot examples before anything user-specific. Cached tokens also skip your rate limits, which matters most on the Free plan’s 8,000 tokens per minute.
Use Flex for bursty jobs
On the Developer plan, setting service_tier to flex in the request body gives roughly 10x the on-demand rate limits at the same per-token price. The trade is that over-capacity requests fail fast with a 498 capacity_exceeded error, so wrap calls in jittered retries and keep latency-sensitive traffic on the default tier.
Transcribe with Whisper turbo
whisper-large-v3-turbo costs $0.04 per audio hour against $0.111 for whisper-large-v3. Ten hours of meeting audio a day is about $12 a month on turbo. The Free plan allows 7,200 audio seconds per hour, so a 2 hour file needs to be chunked or run on the Developer plan.
Set a spend limit before shipping
Spend limits are available only on paid accounts. Under Settings, Billing, Limits click Add Limit and enter a monthly USD amount. Spend tracking lags 10 to 15 minutes, so set the cap slightly under your real ceiling. When it trips, new requests return a 400 blocked_api_access error until you raise the limit or the month resets.
Groq billing, explained
The Free plan never charges anything because there’s no payment method on file; requests above the limits return a 429 and you wait for the minute or the day to reset. Upgrading to Developer asks for a credit card, US bank account or SEPA debit account and triggers no charge. Groq bills in arrears, so you pay for tokens already used, and nothing is invoiced until usage reaches at least $0.50.
New Developer accounts are invoiced when lifetime usage crosses $1, $10, $100, $500 and $1,000, with payment pulled automatically at each threshold. After $1,000 billing becomes monthly. Customers with an Indian billing address follow $1, $10 and then a recurring $100 threshold. A failed payment can be retried twice and may lead to suspension if it keeps failing, so keep the card current.
Downgrading to Free is allowed at any time from Settings, Billing, but Groq issues a final invoice for unbilled usage first and your limits drop back to the Free table. Refunds are case by case through support. Enterprise plans add the Performance tier, sold as provisioned throughput with a 99.9% availability SLA and priced by contract rather than per token.
Groq credits FAQ
is the groq api free to use
Yes, within the Free plan limits and without a card. You get 30 requests per minute, 8,000 tokens per minute, 1,000 requests per day and 200,000 tokens per day on gpt-oss-120b, gpt-oss-20b and qwen3.8-27b, plus 7,200 audio seconds per hour on Whisper. Nothing is charged until you upgrade to the Developer plan and add a payment method.
does groq give free credits to new accounts
No. There’s no signup credit, trial balance or promo code. Promotional credits are offered occasionally at hackathons and special events, applied by Groq rather than redeemed by code. The Free plan is the only standing free allowance, and the $10,000 Groq for Startups page now redirects to the homepage.
what are groq free tier rate limits
Per model on the Free plan: gpt-oss-120b, gpt-oss-20b and qwen3.8-27b get 30 RPM, 1,000 RPD, 8,000 TPM and 200,000 TPD. gpt-oss-safeguard-20b gets 3 RPM and 2,000 TPM. whisper-large-v3 and v3-turbo get 20 RPM, 2,000 RPD, 7,200 audio seconds per hour and 28,800 per day. Orpheus text to speech gets 10 RPM and 100 RPD. Your organization’s limits page is the one that counts.
how much does groq charge per million tokens
gpt-oss-120b is $0.15 input and $0.60 output per million tokens, gpt-oss-20b and gpt-oss-safeguard-20b are $0.075 and $0.30, and qwen3.8-27b is $0.80 and $4.00. Llama 3.3 70B and Llama 3.1 8B are listed as contact sales. Whisper large v3 is $0.111 per audio hour and v3 turbo is $0.04. Batch jobs are half those rates.
does groq have a batch api discount
Yes, 50% off synchronous pricing, with a completion window you set between 24 hours and 7 days and rate limits separate from your per-model limits. Batch requires the Developer plan, which lists batch processing as an included feature. The batch discount doesn’t stack with the 50% prompt caching discount, so pick one per job.
does groq have a startup program
Not one you can apply to right now. The Groq for Startups page, which once described $10,000 in inference credits for hand-selected startups, redirects to the Groq homepage, and the application page does the same. Groq has published no closure notice, so email Groq directly if a VC partner refers you.
when does groq charge my card
In arrears. Nothing is charged at upgrade. Groq invoices and pulls payment automatically when lifetime usage crosses $1, $10, $100, $500 and $1,000, then on a monthly cycle, and never for totals under $0.50. You can see invoices and update the card under Settings, Billing, Manage Billing.
FREE TIER200,000 free tokens per day on gpt-oss-120b
