---
title: "Claude Opus 4.6"
description: "Claude Opus 4.6 improves on Opus 4.5 with stronger coding, planning, debugging, code review, long-running agentic work, and performance in large codebases. It is the first Opus-class model with a beta 1M-token context window, while API features include adaptive thinking, four effort levels, context compaction, and up to 128k output tokens. In reported evaluations, it achieves the highest score on Terminal-Bench 2.0, leads frontier models on Humanity’s Last Exam and BrowseComp, and exceeds GPT-5.2 by around 144 Elo points and Opus 4.5 by 190 points on GDPval-AA. The release also adds Claude Code agent teams, upgraded Excel capabilities, a PowerPoint research preview, and availability on claude.ai, the API, and major cloud platforms, with base pricing of $5/$25 per million tokens."
---

# Claude Opus 4.6

[Anthropic](https://yomu.fyi/company/anthropic) · Jul 22, 2026

**Type:** Announcement

## Summary

Claude Opus 4.6 improves on Opus 4.5 with stronger coding, planning, debugging, code review, long-running agentic work, and performance in large codebases. It is the first Opus-class model with a beta 1M-token context window, while API features include adaptive thinking, four effort levels, context compaction, and up to 128k output tokens. In reported evaluations, it achieves the highest score on Terminal-Bench 2.0, leads frontier models on Humanity’s Last Exam and BrowseComp, and exceeds GPT-5.2 by around 144 Elo points and Opus 4.5 by 190 points on GDPval-AA. The release also adds Claude Code agent teams, upgraded Excel capabilities, a PowerPoint research preview, and availability on claude.ai, the API, and major cloud platforms, with base pricing of $5/$25 per million tokens.

## Context

Claude Opus 4.6 is introduced as an upgrade to Opus 4.5, targeting harder coding tasks, longer-running agentic work, larger codebases, everyday knowledge work, and complex research workflows.

## Approach / What changed

The release combines model improvements with a beta 1M-token context window, adaptive thinking, configurable effort levels, context compaction, 128k output support, Claude Code agent teams, upgraded Excel functionality, and a PowerPoint research preview.

## Takeaways

- Opus 4.6 supports four API effort levels—low, medium, high, and max—and adaptive thinking lets it decide when extended reasoning is useful.
- Context compaction beta automatically summarizes and replaces older context near a configurable threshold, helping long-running conversations and agentic tasks continue without hitting limits.
- On GDPval-AA, Opus 4.6 outperforms GPT-5.2 by around 144 Elo points and Claude Opus 4.5 by 190 points; base pricing remains $5/$25 per million tokens.

**Tags:** [AI Agents](https://yomu.fyi/topic/ai-agents), [Claude](https://yomu.fyi/topic/claude), [LLMs](https://yomu.fyi/topic/llm), [Model Evaluation](https://yomu.fyi/topic/model-evaluation)

- Source: [Anthropic](https://www.anthropic.com/news/claude-opus-4-6)
- Source URL: https://www.anthropic.com/news/claude-opus-4-6
- Ingested by Yomu: 2026-09-09T11:37:16.948Z

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