Anthropic has released Claude Sonnet 4.5, claiming it's the world's best AI model for programming. The new system demonstrates remarkable capabilities in extended coding tasks.
Key Performance Highlights
The model showed impressive endurance, working continuously on the same project for over 30 hours on complex, multi-step tasks. However, Anthropic hasn't disclosed specific details about these projects.
Benchmark Results
Claude Sonnet 4.5 dominated the SWE-bench Verified benchmark with strong results:
- Claude Sonnet 4.5: 82%
- Claude Opus 4.1: 79.4%
- GPT-5 Codex: 74.5%
- Gemini 2.5 Pro: 67.2%
Availability
The tool is now available globally for developers and users worldwide.
Source: Ars Technica
