Archive — US — techcrunch
Anthropic releases Sonnet 4.6
Anthropic has launched Sonnet 4.6, enhancing coding and instruction-following capabilities, featuring a 1 million-token context window. This upgrade is part of Anthropic's routine four-month updates and includes record benchmark scores, positioning it notably within the AI landscape, though still behind some leading models.
Key points
- Release of Sonnet 4.6 highlights improvements in coding and instruction-following.
- Context window expanded to 1 million tokens, enabling comprehensive data handling.
- New model becomes the default for Free and Pro plan users.
- Record benchmark scores achieved, notably a 60.4% on ARC-AGI-2.
- The launch follows Opus 4.6 and precedes an updated Haiku model.
Why it matters
- This release aligns with 2025 IT trends of increasing AI model capabilities.
- The rise in context window sizes signifies a shift towards models capable of processing larger datasets, a notable requirement in the data-rich environments of 2025.
- Competition among AI models is intensifying, as seen with Sonnet 4.6 trailing behind models like Opus 4.6 and GPT 5.2.
Anthropic's Sonnet 4.6 marks a significant advancement in AI capabilities, setting new performance benchmarks, while also highlighting the competitive landscape of AI model development as it strives to catch up with industry leaders.