Anthropic Releases Claude Sonnet 4.6: 1M Token Context, 5x Better Computer Use

Full-Stack Upgrade Across Coding, Agents, and Computer Use

Anthropic released Claude Sonnet 4.6 on February 17, 2026, delivering a comprehensive upgrade across coding, computer use, long-context reasoning, and agentic planning. The model is now the default for both free and paid Claude users worldwide.

Computer Use: Near Human-Level Performance

The headline improvement is in computer use. Claude Sonnet 4.6 achieves 72.5% on the OSWorld benchmark, representing a nearly fivefold improvement since the first computer use release 16 months ago. The model can now navigate complex spreadsheets, fill out multi-step web forms, and operate across multiple browser tabs with what Anthropic describes as "human-level capability."

Coding: Preferred Over Its Own Flagship

In coding evaluations, developers using Claude Code preferred Sonnet 4.6 over its predecessor roughly 70% of the time. More striking: users favored Sonnet 4.6 over Opus 4.5 in 59% of comparisons. The model shows significantly less overengineering and better instruction-following than previous versions.

1M Token Context Window in Beta

Sonnet 4.6 supports a 1 million token context window in beta, enabling processing of entire codebases, lengthy contracts, or dozens of research papers in a single request. The standard context window remains 200K tokens, with 64K maximum output tokens.

Available Across Platforms at Unchanged Pricing

Claude Sonnet 4.6 launched simultaneously in GitHub Copilot (Pro, Pro+, Business, Enterprise) and Amazon Bedrock, in addition to Anthropic's API. Pricing holds steady at $3/$15 per million tokens (input/output), the same as Sonnet 4.5. VentureBeat noted the model "matches flagship AI performance at one-fifth the cost."

Source: Introducing Claude Sonnet 4.6 — Anthropic