Claude Sonnet 4.6: Anthropic's Most Capable Sonnet Yet Changes the AI Game
Anthropic has launched Claude Sonnet 4.6, a major upgrade to its most popular AI model that delivers significant improvements across coding, computer use, long-context reasoning, and agentic tasks. The new model is now the default on free and Pro plans in claude.ai and Claude Cowork, and it's already winning over developers and enterprises alike.
What Makes Sonnet 4.6 Special
The numbers tell the story: in internal testing, users preferred Sonnet 4.6 over its predecessor Sonnet 4.5 approximately 70% of the time in Claude Code. Perhaps even more impressively, users preferred it to the much larger Opus 4.5 model 59% of the time — a remarkable achievement for a smaller, faster model.
Key improvements include:
- 1M token context window now available in beta — that's roughly 750,000 words of context
- Major computer use improvements with significant gains on the OSWorld benchmark
- Better prompt injection resistance — critical for enterprise deployments
- Same pricing as Sonnet 4.5 — $3/$15 per million input/output tokens
- Supports adaptive thinking, extended thinking, and context compaction
This is Claude Sonnet 4.6: our most capable Sonnet model yet.
— Claude (@claudeai) February 17, 2026
It’s a full upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design.
It also features a 1M token context window in beta. pic.twitter.com/TDId3XUSRs
Why Developers Are Excited
The model has received enthusiastic responses from major players in the AI development space. Cursor reported that Sonnet 4.6 is their most popular model. GitHub said it powers Copilot across coding, code review, and multi-file editing. Replit noted major improvements in agentic coding tasks. Cognition highlighted its superior performance in their autonomous software engineering agent Devin.
On the enterprise side, Databricks praised its reasoning capabilities for data engineering workflows. Box said it delivers more sophisticated document intelligence. Rakuten reported improved customer service automation, and Zapier noted better multi-step workflow automation.
The Vending-Bench Arena Surprise
In an interesting benchmark test called Vending-Bench Arena — where AI models compete in a simulated vending machine business — Sonnet 4.6 developed a sophisticated strategy of investing heavily early to build market share, then pivoting to profitability. This kind of strategic reasoning demonstrates the model's improved ability to plan and execute complex, multi-step tasks.
The Bottom Line
Claude Sonnet 4.6 represents a significant step forward in the AI capabilities race. By delivering Opus-level performance at Sonnet pricing, Anthropic is effectively making their most advanced AI capabilities accessible to a much wider audience. The model ID is claude-sonnet-4-6, and it's available now across all Anthropic platforms.
For developers and enterprises already using Claude, the upgrade is seamless — and for those who haven't tried it yet, there's never been a better time to start.