Back
AnthropicClaude API
We've launched Claude Opus 4.6, our most intelligent model for complex agentic tasks and long-horizon work
AI summary
Written by AI from the official notes. Check them for exact details.Claude Opus 4.6 has been launched with new features for complex tasks and improved context handling.
- Adaptive thinking is now recommended; manual thinking is deprecated.
- Effort parameter no longer needs a beta header for Opus 4.6.
- Compaction API in beta for server-side context summarization.
- Data residency controls allow specifying inference locations.
- 1M token context window available in beta for Opus 4.6.
Why it matters: Users needing advanced AI capabilities for complex tasks should explore the new features in Claude Opus 4.6.
Full release notes6 changes
- We've launched Claude Opus 4.6, our most intelligent model for complex agentic tasks and long-horizon work. Opus 4.6 recommends adaptive thinking (
thinking: {type: "adaptive"}); manual thinking (type: "enabled"withbudget_tokens) is deprecated. Opus 4.6 does not support prefilling assistant messages. Learn more in What's new in Claude 4.6. - The effort parameter no longer requires a beta header and now supports Claude Opus 4.6. Effort replaces
budget_tokensfor controlling thinking depth on new models. - We've launched the compaction API in beta, providing server-side context summarization for effectively infinite conversations. Available on Opus 4.6.
- We've introduced data residency controls, allowing you to specify where model inference runs with the
inference_geoparameter. US-only inference is available at 1.1x pricing for models released after February 1, 2026. - The 1M token context window is now available in beta for Claude Opus 4.6, in addition to Sonnet 4.5 and Sonnet 4. Long context pricing applies to requests exceeding 200k input tokens.
- Fine-grained tool streaming no longer requires a beta header on any model or platform.