Claude Sonnet 4.5 is now available, a strong model for coding and for building complex agents.
Coding. State-of-the-art results on real-world software development tasks.
Extended focus. Can stay on complex, multi-step tasks for more than 30 hours.
Computer use. Leading performance on computer interaction tasks, with 61.4% on OSWorld.
Reasoning. Substantial gains in reasoning and math.
Alignment. Fewer concerning behaviors than earlier frontier models.
Grok 4 is also available, adding another option for different tasks and conversational styles.