---
title: "Introducing Sonnet 4.6"
description: "Claude Sonnet 4.6 is introduced as an upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design. It adds a 1M token context window in beta, becomes the default for Free and Pro users in claude.ai and Claude Cowork, and keeps Sonnet 4.5’s starting price of $3/$15 per million tokens. Early testing reports that users preferred it over Sonnet 4.5 in Claude Code roughly 70% of the time and over Opus 4.5 59% of the time. The release also reports gains on OSWorld computer use, prompt-injection resistance, OfficeQA document comprehension, and long-horizon planning, while adding adaptive and extended thinking, beta context compaction, and generally available code execution, memory, programmatic tool calling, tool search, and tool-use examples."
---

# Introducing Sonnet 4.6

[Anthropic](https://yomu.fyi/company/anthropic) · Jul 21, 2026

**Type:** Announcement

## Summary

Claude Sonnet 4.6 is introduced as an upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design. It adds a 1M token context window in beta, becomes the default for Free and Pro users in claude.ai and Claude Cowork, and keeps Sonnet 4.5’s starting price of $3/$15 per million tokens. Early testing reports that users preferred it over Sonnet 4.5 in Claude Code roughly 70% of the time and over Opus 4.5 59% of the time. The release also reports gains on OSWorld computer use, prompt-injection resistance, OfficeQA document comprehension, and long-horizon planning, while adding adaptive and extended thinking, beta context compaction, and generally available code execution, memory, programmatic tool calling, tool search, and tool-use examples.

## Context

Organizations often have specialized software built before modern APIs, making automation difficult and requiring bespoke connectors. Computer use is presented as a way to let models operate such software through interfaces similar to those used by people.

## Approach / What changed

Sonnet 4.6 upgrades model capabilities across coding, computer use, reasoning, planning, knowledge work, and design. The release adds a beta 1M token context window, adaptive and extended thinking, beta context compaction, and several generally available API capabilities, while improving resistance to prompt injection and performance on evaluation benchmarks.

## Takeaways

- Sonnet 4.6 matches Opus 4.6 performance on OfficeQA, which tests extracting and reasoning over facts in enterprise documents, including charts, PDFs, and tables.
- In Vending-Bench Arena, Sonnet 4.6 invested heavily in capacity during the first ten simulated months before pivoting toward profitability, finishing well ahead of competitors.
- The API’s web search and fetch tools can automatically write and execute code to filter search results, retaining relevant content to improve response quality and token efficiency.

**Tags:** [AI Agents](https://yomu.fyi/topic/ai-agents), [Claude](https://yomu.fyi/topic/claude), [Computer Use](https://yomu.fyi/topic/computer-use), [LLMs](https://yomu.fyi/topic/llm)

- Source: [Anthropic](https://www.anthropic.com/news/claude-sonnet-4-6)
- Source URL: https://www.anthropic.com/news/claude-sonnet-4-6
- Ingested by Yomu: 2026-09-09T11:37:19.648Z

[Read original post](https://www.anthropic.com/news/claude-sonnet-4-6)