Anthropic Launches Claude Opus 5.5 at 40% Lower Cost, Boosting Speed and Coding Performance
Updated
Updated · KDnuggets · Sep 23
Anthropic Launches Claude Opus 5.5 at 40% Lower Cost, Boosting Speed and Coding Performance
3 articles · Updated · KDnuggets · Sep 23
Summary
Claude Opus 5.5 went live Sept. 22 as Anthropic’s first 5.5-family model, positioned near Claude Fable 5.1-level performance while typically costing 40% less than Opus 5 and generating output more than 30% faster.
Independent and internal tests backed the efficiency push: Artificial Analysis scored it at 58, output speed reached 74 to 86 tokens per second, and Anthropic said a HAProxy C-to-Rust translation finished in 9.5 hours versus 12 for Fable 5.1 at 51% lower cost.
Enterprise examples centered on coding and knowledge work, including a 680,000-line migration completed in under a day, a 200,000-line codebase audit in under three hours, and 16 of 18 successful earnings-report attempts where Opus 5 and Fable 5.1 cleared none.
The release also expands access and changes deployment terms, with a 1 million-token context window, 128,000-token max output, Batch API pricing discounted 50%, Fast mode up to 2.5x quicker, and support across Anthropic’s API, Bedrock, Google Cloud and Microsoft Foundry.
Anthropic framed Opus 5.5 as a slower, safety-led launch: internal audits showed about 85% fewer containment-crossing attempts than Opus 5 or Mythos 5.1, though the system card disclosed weaker resistance to malicious pasted instructions and unverified authorization claims.