Claude Haiku 5.5
Claude Haiku 5.5 is built for repeated classification, extraction and routing tasks. Its headline price applies to shorter prompts; longer requests have a different rate.
Compare actual token counts when upgrading
Haiku 5.5 uses a newer tokenizer than Haiku 4.5. A rate comparison based on the older model’s token count can therefore misstate the cost, even before a request crosses the long-prompt pricing threshold.Claude Haiku 5.5 reference
Moving a conversation can affect reasoning history
Thinking blocks are restricted to the original account or linked accounts. An application that moves work between accounts needs to account for that boundary.Claude Haiku 5.5 reference
Make missing information a valid result
Practical guidance. For invoice extraction, define the required fields and allow an explicit missing value. Review currency, tax and totals separately. A cheap extraction pipeline becomes less useful if it silently fills gaps with plausible numbers.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | |
| Access | API |
| Model identifiers | claude-haiku-5-5 |
| Context window | 1,000,000 tokens Context limit Claude Haiku 5.5 reference |
| Maximum output | 128,000 tokens Ordinary maximum output Claude Haiku 5.5 reference |
| Input price | 0.1 USD per million tokens Prompts up to 100,000 tokens Claude Haiku 5.5 reference |
| Output price | 0.5 USD per million tokens Prompts up to 100,000 tokens Claude Haiku 5.5 reference |
| Input price | 0.5 USD per million tokens Prompts over 100,000 tokens Claude Haiku 5.5 reference |
| Output price | 2.5 USD per million tokens Prompts over 100,000 tokens Claude Haiku 5.5 reference |
| Thinking records | Usable only in the producing account or an account linked to it.Claude Haiku 5.5 reference |
| Tokenizer change | The same text can count as about 30% more tokens than with Haiku 4.5.Claude Haiku 5.5 reference |
Related models
- Claude Sonnet 5.5
Anthropic · Claude Sonnet
A model for coding, writing and agent tasks with configurable reasoning.
Launch date API - Claude Haiku 4.5
Anthropic · Claude Haiku
An earlier Haiku model with manual reasoning budgets.
Launch date API