Claude Sonnet 4.6

• 🆕 Anthropic's mid-tier Sonnet upgrade, released February 2026.
• 📏 1M-token context window with 64K max output tokens.
• 🔧 Improved coding, reasoning, and general-task performance.
• 👁️ Multimodal with vision, function calling, and web search.
• 🎯 Tuned for a strong quality-to-cost balance.
• 💬 Anthropic reports 80.2% on SWE-bench Verified (with prompt modification).
• 🧠 Positioned below Anthropic's flagship Opus tier.
• 🏢 Available through the Claude API.

Anthropic

Anthropic is an American artificial intelligence company headquartered in San Francisco, focused on developing large language models under the Claude family. Founded with a strong emphasis on AI safety research, the company has become one of the most…

Claude Sonnet 4.6 is the latest entry in Anthropic's mid-tier Sonnet line, positioned as a balance of speed, intelligence, and cost for coding, reasoning, and general production workloads. Anthropic describes it as a full upgrade across coding, reasoning, and general tasks, paired with a 1-million-token context window and up to 64K output tokens.

Relative to its same-family predecessor Claude Sonnet 4.5, the most concrete change is the context window, which expands to 1M tokens, letting the model ingest entire codebases, lengthy documents, or large research bundles in a single request. Anthropic positions the release as an across-the-board improvement over the prior Sonnet generation in coding and reasoning quality.

On agentic coding, Anthropic reports an 80.2% score on SWE-bench Verified with a prompt modification, and presents the model as narrowing the gap with its heavier Opus tier on demanding tasks while retaining the cost efficiency that defines the Sonnet line.

For the most complex, high-autonomy work, Anthropic still points users toward its flagship Opus tier, such as Claude Opus 4.8. Claude Sonnet 4.6 supports vision, function calling, and web search, and is accessible through the Claude API.