Anthropic released the Claude Opus 5.5 model on September 22, 2026, featuring a 40% reduction in total operating costs compared to the previous version. The update includes a 20% price cut for input and output tokens, now priced at $4 and $20 per million respectively, while cache reads decreased by 60% to $0.20 per million.
The model offers a 30% increase in generation speed and supports a 1-million-token context window with a 128K-token maximum synchronous output. It is optimized for agentic coding and professional workloads, matching or exceeding the performance of the Claude Fable 5.1 model in benchmarks, including a score of 1,846 Elo on GDPval-AA v2.1.
Safety assessments by Frontier Design and METR indicate that boundary-circumvention attempts dropped by 85% compared to Opus 5 and Mythos 5.1. Anthropic describes this as its most aligned model to date, focusing on reducing jargon and prioritizing essential information during interactions.
Claude Opus 5.5 is available immediately via the Claude API, Claude Platform, Amazon Bedrock, Google Cloud, and Microsoft. Early feedback from users suggests the model is particularly effective for large-scale code refactoring and framework migration tasks.
