Limits and plans
Every organization has a plan. Your plan, its limits and up-to-date usage are in the console.
What is counted
- Uncached input tokens. When you continue a conversation the model has already seen, the repeated part is read from the cache and doesn’t count: it costs almost nothing.
- Output tokens, reasoning included.
The beta plans count the two numbers apart; the public plans (Free, Plus, Pro, Max) count their sum.
The windows
| Window | How it works |
|---|---|
| Last 5 hours | Rolling: usage leaves the count 5 hours after the request |
| Last 7 days | Rolling, like the 5 hours (Free plan) |
| Month | Calendar month, Italian time: it resets at midnight on the 1st |
| Period | Paid plans: 30 days from activation, then it renews |
| Concurrency | Requests running or queued at the same time for your organization |
The check happens before every request. A request that starts under the limit runs to the end, even if it goes over.
When you reach a limit
- 5-hour window, 7 days or too many requests at once:
429 rate_limit_exceeded, withRetry-After. - Month or period:
429 insufficient_quota, withRetry-Afteruntil the 1st of the month or the renewal. - For quotas we add
x-should-retry: false, so the SDKs don’t insist for nothing.
Agentic tools like Cline or OpenCode sometimes send several requests at once. If your plan allows one at a time, set them to work sequentially.
Plans
During the beta we assign plans. The public plans are announced and not on sale yet (prices include VAT, renewal every 30 days):
| Plan | Price a month | Tokens every 5 hours | Other limits | Requests at once |
|---|---|---|---|---|
| Free | €0 | 30,000 | 150,000 every 7 days; Chat only, no API; reasoning off; basic web search | 1 |
| Plus | €19 | 500,000 | 2,000,000 every 30 days | 1 |
| Pro | €89 | 5 times Plus | monthly usage limits apply | 2 |
| Max | €169 | 20 times Plus | monthly usage limits apply | 2 |
The Free plan includes the Chat only: no API keys, and a key of an organization moved to Free gets 403 api_not_in_plan. One request reads at most about 30,000 tokens (400 request_too_large). The API comes with the Plus plan.
On the Free plan the model always answers without reasoning: reasoning_effort is ignored and counts as none. In the chat, web search is basic: up to 3 searches or pages read per answer, 5 results per search, 10 searches and 20 pages an hour. From the Plus plan reasoning has all four levels and web search is complete.
beatrice-flash-extreme costs 50% more on the paid plans, for verified organizations (see Models). If you need more capacity, write to us.