Hybrid AI reasoning model

Anthropic launches Claude 3.7 Sonnet, introducing hybrid AI reasoning.

The model scored 62.3% on the SWE-Bench test for coding tasks, outperforming OpenAI's o3-mini at 49.3%.