---
meta:
title: Pricing and rate limits
description: Pay-as-you-go pricing and the free and paid rate-limit tiers for Meta Model API.
keywords: pricing, pay-as-you-go, per-token pricing, cost, rate limits, requests per minute, tokens per minute, throttling, background
cms:
alias: /model-api/docs/getting-started/pricing-rate-limits
target: aidmc
---
# Pricing and rate limits
## Pricing {#pricing}
You pay only for what you use. Meta Model API bills per token with no minimums or upfront commitment, priced per 1M tokens in USD.
| Usage | Price per 1M tokens |
| :---- | :---- |
| Input | $1.25 |
| Cached input | $0.15 |
| Output | $4.25 |
Cached input costs less. When part of your prompt matches a [cached prefix](/docs/features/prompt-caching), you pay the cached-input rate for those tokens. Check `cached_tokens` in the response to see how many were served from cache.
**Web search grounding** costs **$2.50 per 1,000 search queries**, in addition to tokens for the request. See [Search grounding](/docs/features/search-grounding).
There is **no long-context premium**: you pay the same rate whether your context window is mostly empty or almost full.
## Rate limits {#rate-limits}
Rate limits control per-minute usage for your team. Two tiers apply:
| Tier | Requests per minute (RPM) | Tokens per minute (TPM) |
| :---- | :---- | :---- |
| Free | 60 | 2,000,000 |
| Paid | 3,000 | 4,000,000 |
Limits apply **per team, not per API key**. If you use multiple keys in one team, all requests and tokens count toward one shared quota.
Token usage counts only **your** input and output tokens. Meta injects a small amount of steering context into every prompt (a system prompt and related scaffolding); those injected tokens are not billed and are excluded from the token counts reported in your `usage` and by the [token-counting endpoints](/docs/features/token-counting#injected-tokens).
## Retry on 429 responses {#rate-limit-behavior}
If you exceed RPM or TPM, the API returns `HTTP 429 Too Many Requests`. Requests succeed again once usage drops below the limit.
**Retry strategy:** Use exponential backoff with jitter. Start with a short delay such as 500 ms, double it after each consecutive `429`, and add random jitter to avoid thundering-herd effects across concurrent clients.
## Read rate-limit headers {#rate-limit-headers}
Every successful response includes headers that report your current limit and remaining quota:
| Header | Description |
| :---- | :---- |
| `x-ratelimit-limit-tokens` | Total token budget for the current window. |
| `x-ratelimit-remaining-tokens` | Tokens remaining before you hit the limit. |
| `x-ratelimit-limit-requests` | Total request budget for the current window. |
| `x-ratelimit-remaining-requests` | Requests remaining before you hit the limit. |
Read these headers and slow down before you hit a `429` instead of reacting after one.
## Submit background responses within limits {#background-rate-limit}
Creating [background responses](/docs/features/responses#using-with-other-features) (`background: true`) counts against a separate per-team submission limit in addition to the RPM and TPM limits above. The default is **600 background submissions per minute, per team**.
If you exceed that cap, new background submissions return `HTTP 429 Too Many Requests` with a `Retry-After` header. Responses already running continue unaffected. Apply the same exponential-backoff-with-jitter strategy described above.
The cap applies only to starting new background responses. Standard RPM and TPM limits still apply to every request.
## Next steps
- [Authentication](/docs/getting-started/authentication): create and manage API keys
- [Get started](/docs/getting-started/overview#first-call): make your first API call
- [Chat completion](/docs/features/chat-completion): start generating text with the core conversational endpoint
Pricing and rate limits
You pay only for what you use. Meta Model API bills per token with no minimums or upfront commitment, priced per 1M tokens in USD.Usage | Price per 1M tokens |
|---|
Input | $1.25 |
Cached input | $0.15 |
Output | $4.25 |
cached_tokens in the response to see how many were served from cache.Web search grounding costs $2.50 per 1,000 search queries, in addition to tokens for the request. See Search grounding.There is no long-context premium: you pay the same rate whether your context window is mostly empty or almost full.Rate limits control per-minute usage for your team. Two tiers apply:Tier | Requests per minute (RPM) | Tokens per minute (TPM) |
|---|
Free | 60 | 2,000,000 |
Paid | 3,000 | 4,000,000 |
usage and by the token-counting endpoints.If you exceed RPM or TPM, the API returns HTTP 429 Too Many Requests. Requests succeed again once usage drops below the limit.Retry strategy: Use exponential backoff with jitter. Start with a short delay such as 500 ms, double it after each consecutive 429, and add random jitter to avoid thundering-herd effects across concurrent clients.Every successful response includes headers that report your current limit and remaining quota:Header | Description |
|---|
x-ratelimit-limit-tokens
| Total token budget for the current window. |
x-ratelimit-remaining-tokens
| Tokens remaining before you hit the limit. |
x-ratelimit-limit-requests
| Total request budget for the current window. |
x-ratelimit-remaining-requests
| Requests remaining before you hit the limit. |
429 instead of reacting after one.Submit background responses within limits
background: true) counts against a separate per-team submission limit in addition to the RPM and TPM limits above. The default is 600 background submissions per minute, per team.If you exceed that cap, new background submissions return HTTP 429 Too Many Requests with a Retry-After header. Responses already running continue unaffected. Apply the same exponential-backoff-with-jitter strategy described above.The cap applies only to starting new background responses. Standard RPM and TPM limits still apply to every request.