Current documented limits
| Constraint | Current value | Scope | Verified source |
|---|---|---|---|
| Messages API limitersThe API uses a token bucket, and limits can be enforced over shorter intervals than one minute. | RPM + input TPM + output TPM | Usage-tier and model specific | AnthropicAug 22, 2026 |
| Claude Fable 5, Opus 5, and Sonnet 5 contextAPI model specification; model availability can differ by provider. | 1,000,000 tokens | Claude 5 family | AnthropicAug 22, 2026 |
| Claude 5 family maximum outputSeparate maximum output for the listed Claude 5 API models. | 128,000 tokens | Claude 5 family | AnthropicAug 22, 2026 |
| Claude Haiku 4.5 contextModel-specific API context window. | 200,000 tokens | Claude Haiku 4.5 | AnthropicAug 22, 2026 |
How to apply Anthropic API limits safely
The monitored baseline covers Messages API limiters, Claude Fable 5, Opus 5, and Sonnet 5 context, Claude 5 family maximum output, Claude Haiku 4.5 context. Treat these as separate constraints rather than one platform-wide capacity number: a workload can fit one row and still fail another because the plan, model, endpoint, runtime, region, invocation mode, or account scope differs.
- Match the production workload to the exact scope printed beside each value and confirm it in the active Anthropic API console, configuration, or response headers.
- Measure the serialized request, token volume, duration, concurrency, storage, or connection demand at realistic percentiles, then preserve headroom for bursts and retries.
- Check every adjacent layer—client, SDK, gateway, proxy, queue, database, and downstream service—for a smaller effective limit before changing architecture.
Specific limit pages
These pages exist because the constraint has a distinct implementation or troubleshooting intent. Closely related keyword variations stay consolidated.
Anthropic API rate limits
How Claude API organization, workspace, model, and token-bucket rate limits work, with current official-source guidance.
Claude context windows
Current Claude API context-window and output constraints by model family, with official model documentation and practical caveats.
Errors linked to these limits
529 — overloaded_error
Anthropic API overloaded_error: direct meaning, scoped diagnosis, corrective action, official source, and linked current limit.
429 — rate_limit_error
Diagnose Anthropic API 429 rate_limit_error responses across request, input-token, output-token, and acceleration controls.
Compare alternatives
Gemini API vs Anthropic API Limits
Gemini API vs Anthropic API Limits using the same verified structured observations as each underlying limit page, with scope and comparability caveats.
OpenAI API vs Anthropic API
Compare current context-window and rate-limit models while preserving model, tier, project, and organization caveats.
ChatGPT vs Claude
A cautious comparison of disclosed ChatGPT product upload constraints and Claude API model context specifications without inventing parity.