What happened
Anthropic released Claude Sonnet 4.6, an upgraded mid-size AI model, on February 17, 2026. The model introduces a 1 million token context window in beta — doubling previous capacity — alongside significant improvements in coding, computer use, and agent planning. Sonnet 4.6 is now the default model for free and Pro plan users on claude.ai and Claude Cowork, maintaining existing pricing at $3 per million input tokens and $15 per million output tokens.
The release arrives 12 days after Claude Opus 4.6, reinforcing Anthropic's four-month update cycle across its model family. On benchmarks, Sonnet 4.6 scored 72.5% on OSWorld-Verified for computer use — up from 28.0% by Sonnet 3.7 a year earlier — and set new records on SWE-Bench for software engineering. In internal Claude Code testing, users preferred Sonnet 4.6 over Sonnet 4.5 approximately 70% of the time. An updated Haiku model is expected to follow in coming weeks.
Why it matters
CTOs and platform engineers face accelerated model evaluation cycles as Anthropic maintains a predictable update cadence across its Opus, Sonnet, and Haiku tiers. Sonnet 4.6's near-flagship performance at mid-tier pricing reshapes the cost-capability equation — performance that previously required Opus-class models is now available at Sonnet pricing. For procurement teams, this intensifies competitive pressure across the enterprise AI market. The 1M token context window positions Sonnet 4.6 for large-scale document processing and codebase analysis that previously required more expensive models.
Subscribe for Weekly Updates
Stay ahead with our weekly AI and tech briefings, delivered every Tuesday.
