Claude Haiku 5.5 launches with a 1M-token context; Sonnet 5.5 cache reads cut 50%
Anthropic's new small model is built for fast, high-volume work. Sonnet 5.5 also gets cheaper cache reads, and Max/Team plans now include monthly API credits.
Claude Haiku 5.5 (claude-haiku-5-5) is Anthropic’s new model for latency-sensitive, high-volume work. It has a 1M-token context window, up to 128k output tokens, and adaptive thinking controlled by the effort parameter. It’s available on the Claude API, Amazon Bedrock, Google Cloud and Microsoft Foundry.
Alongside it:
- Sonnet 5.5 prompt-cache reads drop from $0.20 to $0.10 per million tokens.
- Monthly API credits for subscribers: $100 (Max 5x), $200 (Max 20x), and up to $500 pooled for Team plans.
- The Python and TypeScript SDKs add computer use and browser use (beta).
Migrating? The release notes flag thinking-related breaking changes, so read the migration guide first.
Why it matters
Cheap, fast models are what you run subagents and bulk jobs on. The Sonnet cache-read cut (Anthropic estimates ~20% cheaper on most agentic work) and the free monthly API credits lower the cost of building with Claude.