claude-sonnet-4-0 — pricing, context window and limits
Retired on the Claude API on 2026-06-15; Amazon Bedrock and Google Cloud still list it on their own schedules. Migrate to Claude Sonnet 5.5.
Price per million tokens
| Input | $3 |
|---|---|
| Output | $15 |
| Cache write, 5-minute TTL | $3.75 |
| Cache write, 1-hour TTL | $6 |
| Cache read | $0.30 (0.1x input) |
| Batch input / output | $1.50 / $7.50 |
Limits and behaviour
| Status | Retired on the Claude API |
|---|---|
| Context window | 200K tokens |
| Max output | 64K tokens |
| Min cacheable prefix | 1,024 tokens |
| Effort levels | not supported |
| Effort default | — |
| Thinking | Fixed budget_tokens, off unless requested |
| Forced tool_choice | Accepted |
| Platforms | Amazon Bedrock, Google Cloud |
Lifecycle
| API model name | claude-sonnet-4-20250514 |
|---|---|
| Deprecated | 14 Apr 2026 |
| Retired | 15 Jun 2026 |
| Still listed on | Amazon Bedrock, Google Cloud |
| Successor | Claude Sonnet 5.5 |
Retired on the Claude API
Anthropic has retired Sonnet 4 from the Claude API; the dates it was deprecated and removed are in the lifecycle table above. This page used to describe it as deprecated with no retirement date announced; Anthropic announced one, and the model is now gone from the first-party API. A request to this id there fails.
Still listed on two partner platforms
Amazon Bedrock and Google Cloud keep their own retirement schedules, and Anthropic's pricing page still lists this model on both. That is not a reason to keep building on it: the schedule belongs to someone else, and this site does not track it.
No effort parameter on this generation
Thinking control here runs through the older budget-based mechanism, which doesn't map directly onto the effort levels its successors use. See migrating off budget_tokens to the effort parameter for that transition.
Auditing where this id might still be hiding
A model this far back is more likely to sit in an older integration nobody has touched recently than in closely watched code. Search the codebase and your configuration for the id explicitly; a retired model can hide in a fallback list that only fires on errors, where nobody notices it failing.
Related
See claude-sonnet-5-5 for the current Sonnet, and the model ID you're calling has been retired for the failure and the fix.
Verified 2026-09-30 against claude-api skill — shared/models.md, model-migration.md + SKILL.md model table and Anthropic's pricing and model-deprecation pages (canonical: https://platform.claude.com/docs/en/about-claude/models/overview, https://platform.claude.com/docs/en/about-claude/pricing, https://platform.claude.com/docs/en/about-claude/model-deprecations).Could not confirm: Partner platforms keep their own retirement dates, which we do not track — check them before relying on this id there.
Verified 2026-10-28 against Anthropic's model-deprecations status table and pricing page (canonical: https://platform.claude.com/docs/en/about-claude/model-deprecations, https://platform.claude.com/docs/en/about-claude/pricing).Could not confirm: These dates apply to the Claude API, Claude Platform on AWS, Claude Models on Google Cloud (Anthropic managed) and Microsoft Foundry. Amazon Bedrock and Claude Models on Google Cloud (Google managed) keep their own retirement schedules, which we do not track.