Claude Sonnet 5.5 keeps Sonnet's price and changes the thinking-off setting

Anthropic released Claude Sonnet 5.5 on September 28, six days after Opus 5.5. The API model is claude-sonnet-5-5.

What it costs

At launch, Sonnet 5.5 kept Sonnet 5's price: $2 input, $10 output, $0.20 cache reads, and $2.50 cache writes per million tokens. Anthropic says it typically needs fewer tokens and costs up to 30% less per task than Sonnet 5, in its own testing.

On October 7, Anthropic halved Sonnet 5.5 cache reads to $0.10 per million tokens and said that makes it around 20% cheaper on most agentic work.

What changed

• Anthropic says output is 30%+ faster than Sonnet 5.
• On Anthropic's table, Terminal-Bench 4.0 is 70.6% against Sonnet 5's 10.3%. The New Stack notes the benchmark's maintainers said Sonnet 5 sometimes hit timeouts and token limits, which helps explain its low score.
• It is the first Sonnet to launch with cyber safeguards. Higher-risk security tasks visibly fall back to Sonnet 5.
• Default effort is Medium in Claude Code and the apps, and High on the Claude Platform.

Before you migrate

If you run Sonnet with thinking off, Anthropic says you need to switch to the new between_tools setting before moving to Sonnet 5.5. Preserved thinking also expands, so thinking cannot be moved between accounts, including when you switch accounts mid-session in Claude Code.

Should you switch

Same list price, with a cheaper cache since October 7. The work is the between_tools change if your integration turns thinking off.