Claude Opus 4.6 released with adaptive thinking and compaction API

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" with budget_tokens) is deprecated. The effort parameter is now generally available and supports Claude Opus 4.6. We've launched the compaction API in beta, providing server-side context summarization for effectively infinite conversations. Introduced data residency controls with the inference_geo parameter for US-only inference at 1.1x pricing. The 1M token context window is now available in beta for Claude Opus 4.6. Fine-grained tool streaming is now generally available on all models.