Guides
Limits & quotas
Codiv is a free public experiment. Limits keep the shared GPU fast for everyone.
Token quotas
Every account gets two free allowances, shown separately on the dashboard:
| Quota | Free | What counts |
|---|---|---|
| System One | 100M tokens | Input tokens, plus thought tokens when you use think |
| Text generation | 10M tokens | Input and output tokens of chat completions |
When a quota is used up, requests to that endpoint return 429: quota_exceeded_error for System One and insufficient_quota for chat completions. The other quota is unaffected. Requests already in flight at that moment still complete. Paid plans with larger quotas are coming soon.
Rate limits
| Limit | Value | When exceeded |
|---|---|---|
| Requests per key | 1,200 per minute (default) | 429 rate_limit_error |
| Active keys per account | 10 | Key creation is refused |
| Server capacity | Shared | 529 overloaded_error with retry-after |
| Text generations at once | 8 across all users, 32 waiting | 529 overloaded_error with retry-after |
Request limits
- Context window: 65,536 tokens. The state and questions are read together, so keep the state a little under that (about 60,000 tokens leaves room for a large question set).
- Request body: 8 MB.
- Images: up to 8 per request, 5 MB each, about 280 tokens each.
- choice: up to 255 options (24 on
verdict-1.4). score: 2 to 10 levels. - Questions per request: no fixed limit. Large sets are answered in parallel chunks.
Retries
Retry 429 rate_limit_error and 529 with exponential backoff, and honour retry-after when it is present. The TypeSafe SDKs do this for you. Do not retry quota_exceeded_error or insufficient_quota.