Claude Sonnet 4.6

Claude Sonnet 4.6, released by Anthropic on February 17, 2026, delivers near-flagship performance across coding, computer use, long-context reasoning, agent planning, knowledge work, and design - all at one-fifth the cost of Opus 4.6. It features a 1 million token context window (beta), matching Opus 4.6's context capacity. In head-to-head comparisons, users preferred Sonnet 4.6 over Sonnet 4.5 in 70% of cases and over the previous flagship Opus 4.5 in 59% of comparisons.

Sonnet 4.6 scores 79.6% on SWE-bench Verified (only 1.2 points behind Opus 4.6's 80.8%), 72.5% on OSWorld-Verified for computer use (nearly matching Opus 4.6's 72.7%), and 61.3% on MCP-Atlas for scaled tool use. On GDPval-AA for office productivity tasks, Sonnet 4.6 reaches 1633 Elo - actually ahead of all models including Opus 4.6 in this specific category. It also scores 63.3% on Finance Agent for financial analysis tasks.

Priced at $3 per million input tokens and $15 per million output tokens (with up to 90% savings via prompt caching and 50% with batch processing), Sonnet 4.6 represents the strongest value proposition in frontier AI. Its combination of near-Opus performance at mid-tier pricing makes it the default choice for most production deployments, particularly for agentic coding, computer use, and knowledge work applications where the marginal performance gap to Opus doesn't justify the 5x price premium.