Anthropic has refreshed its mid-tier and lightweight model lineup with Claude Sonnet 5.5 and Claude Haiku 5.5. Both models feature major speed improvements and enhanced precision when controlling computer desktop interfaces.

Sonnet 5.5 remains priced at $2 per million input tokens and $10 per million output tokens, while Haiku 5.5 is priced at $0.80 input and $4 output. Both models now support a standard 500,000-token context window with sub-second time-to-first-token in supported regions.

The update cements Sonnet 5.5 as the default model in Claude Code and enterprise developer deployments, offering significantly lower error rates on multi-file refactoring and automated UI testing.

What this means

  • Claude 5.5 models double token throughput for computer use tasks.
  • Haiku 5.5 offers near-instant frontier reasoning at $0.80 per million input tokens.
  • Sonnet 5.5 becomes the new default engine across Claude Code and enterprise seats.

Read the announcement

We explained the update in plain English. Use the original source below for the full announcement and any later changes.

Read the original source