What the Llama family is
Meta's Llama models are the default general-purpose choice on most inference platforms, so the same model ID often appears at several providers with very different free-tier limits.
Providers offering Llama on free terms
9 of the 25 providers detailed in this catalog list at least one Llama model. Because Meta publishes the weights rather than the service, the model id can be identical across all of them while the free allowance is not.
| Provider | Free access | Published limits | Card | Matching models |
|---|---|---|---|---|
| GroqCloud | Provider free tier | 30 requests per minute, 1000 requests per day | Not required | llama-3.3-70b-versatile, llama-3.1-8b-instant |
| SambaNova Cloud | Provider free tier | 20 requests per minute, 20 requests per day | Not required | Meta-Llama-3.3-70B-Instruct |
| Cloudflare Workers AI | Provider free tier | Published in compute units | Not required | @cf/meta/llama-3.3-70b-instruct-fp8-fast |
| Fireworks AI | Provider free tier | 10 requests per minute | Not required | accounts/fireworks/models/llama-v3p3-70b-instruct |
| IBM watsonx.ai | Free trial credit | Enforced but not published | Not required | meta-llama/llama-3-3-70b-instruct |
| Together AI | Metered access | Dynamic, no fixed numbers | Not required | meta-llama/Llama-3.3-70B-Instruct-Turbo |
| Nebius Token Factory | Metered access | 60 requests per minute | Not required | meta-llama/Llama-3.3-70B-Instruct |
| DeepInfra | Metered access | Enforced but not published | Not required | meta-llama/Llama-4-Scout-17B-16E |
| Scaleway Generative APIs | Metered access | Enforced but not published | Required | llama-3.3-70b-instruct |
How the free terms actually differ
- GroqCloud — Free plan limits are per model and enforced per organization.
- SambaNova Cloud — The Free Tier applies whenever no payment method is linked to the account.
- Cloudflare Workers AI — Both the Workers Free and Workers Paid plans include 10,000 Neurons per day at no charge, resetting daily at 00:00 UTC.
- Fireworks AI — An account with no payment method and no credits is limited to 10 requests per minute across the entire account.
- IBM watsonx.ai — IBM offers a no-cost trial of watsonx.ai alongside the paid Essentials and Standard plans, and the pricing page invites you to start building at no cost.
- Together AI — Together applies dynamic per-model rate limits that rise with sustained successful traffic and fall when traffic drops, and states plainly that there are no fixed per-model limits published.
- Nebius Token Factory — Limits are dynamic with a published baseline of 60 RPM and 400,000 TPM.
- DeepInfra — The official pricing page lists a per-million-token input and output price for every serving model and does not describe a free allowance or publish request-rate limits.
- Scaleway Generative APIs — Every model served through Generative APIs - Serverless is limited by tokens per minute, queries per minute and concurrent requests.
Picking one
4 of these 9 publish a fixed request count; the rest state a tier or a credit balance instead, which means the only honest answer to "how much is free" is the one in their console. 8 take no credit card. 6 answer a cross-origin browser request, so a key for those can be checked from the browser checker with nothing installed.
A model id that matches Llama is not a promise that the free tier includes it. Providers move individual models between paid and free without changing the id, which is why each row links to the provider page carrying that provider's own wording and the date it was read.