Anthropic releases Claude Sonnet 4.5, tops coding benchmarks
Claude Sonnet 4.5 launched September 29, SWE-bench Verified 77.2%, OSWorld 61.4%, priced at $3/$15 per 1M tokens.
Anthropic released Claude Sonnet 4.5 on September 29, calling it the best coding model in the world. The model scores 77.2% on SWE-bench Verified and leads at 61.4% on the OSWorld computer-use benchmark (Sonnet 4 was at 42.2% four months earlier).
Sonnet 4.5 supports a 200K context window and can sustain complex multi-step tasks for over 30 hours continuously. Anthropic simultaneously launched the Claude Agent SDK, opening the infrastructure behind Claude Code to developers. Claude Code gained checkpoint functionality to save progress and roll back instantly.
Pricing remains the same as Sonnet 4 at $3 input and $15 output per 1M tokens, with up to 90% savings via prompt caching and an additional 50% via batch processing. Sonnet 4.5 is available via the Claude API, Amazon Bedrock, and Google Vertex AI.