Anthropic Launches Claude Opus 5.5 at 40% Lower Cost With Fable-Level Performance Ahead of IPO

Anthropic on Tuesday released Claude Opus 5.5, the first model in its new 5.5 family and its most significant launch since CEO Dario Amodei publicly called for an AI slowdown. The model matches or exceeds the performance of Anthropic's larger Fable 5.1 system on most tasks while costing 40% less to run than Opus 5 and generating output more than 30% faster. Pricing is set at $4 per million input tokens, $20 per million output tokens, and $0.20 per million cache reads, the last figure representing a 60% reduction from Opus 5.

Benchmark gains are substantial across the board. Opus 5.5 scored 66.4% on Terminal-Bench 4.0 versus 52.3% for Opus 5, 57.8% on CursorBench 4.0 versus 46.6%, 1846 on GDPval-AA versus 1708, and 81.8% on OSWorld 2.0 versus 74.0%. In a practical coding test, the model audited and fixed a 200,000-line codebase in under 3 hours compared to more than 20 hours for Opus 5. Anthropic says Opus 5.5 is especially strong on agentic coding, computer use, and long-running knowledge work.

Safety improvements are also highlighted. Anthropic says Opus 5.5 scored best among its recent models on its behavioral alignment audit and attempted to cross containment boundaries approximately 85% less often than either Opus 5 or Mythos 5.1. The model is available immediately through Claude, AWS, Google Cloud, and Microsoft Azure, with Claude Sonnet 5.5 and Haiku 5.5 expected in the coming weeks. The launch comes as Anthropic prepares for a highly anticipated IPO, with Bloomberg noting the release is part of the company's effort to stay ahead of rivals ahead of its Wall Street debut.