What happened

Cognition says Claude Opus 5.5 became available in Devin Desktop and Devin CLI as of September 22, 2026, and Windsurf's blog states it is live there as the new number one on FrontierCode 1.1, edging out Fable 5 and GPT-6 Astra at less than a tenth of the cost of Fable 5.

AWS ML announced Claude Opus 5.5 on Amazon Bedrock and Claude Platform on AWS as the first of the Claude 5.5 family, and The Decoder reports Anthropic positions it as matching Fable 5.1 on most tasks at about forty percent lower run cost than Opus 5, with Sonnet 5.5 and Haiku 5.5 expected soon.

Evidence

• Cognition reports Opus 5.5 is available in Devin Desktop and Devin CLI as of September 22, 2026.

  Windsurf Blog · attributed

  Cognition says Claude Opus 5.5 is now available in Devin Desktop and Devin CLI as of September 22, 2026.

• On FrontierCode 1.1, Cognition reports Opus 5.5 leads Fable 5 and GPT-6 Astra at under one-tenth of Fable 5 cost per task.

  Windsurf Blog · attributed

  On its FrontierCode 1.1 engineering benchmark, Cognition reports Opus 5.5 leads Fable 5 and GPT-6 Astra at under one-tenth of Fable 5 cost per task.

• AWS states Claude Opus 5.5 is available on Amazon Bedrock and Claude Platform on AWS as the first Claude 5.5 family model.

  AWS ML · attributed

  Today, we're excited to announce the availability of Claude Opus 5.5 on Amazon Bedrock and Claude Platform on AWS , the first of the Claude 5.5 model family.

• Anthropic says Opus 5.5 matches Fable 5.1 on most tasks while costing about forty percent less to run than Opus 5.

  The Decoder · attributed

  The company says it matches Claude Fable 5.1 on most tasks while costing about 40 percent less to run than Opus 5.

• Teknium states Claude Opus 5.5 is available in Hermes Agent via Nous Portal.

  Teknium · attributed

  Anthropic's Claude Opus 5.5 is now available in Hermes Agent via Nous Portal!

• Techmeme summarizes The Verge reporting enhanced safeguards and behavior improvements on Opus 5.5.

  Techmeme · attributed

  Anthropic launches Claude Opus 5.5, its first model since Dario Amodei's "pace the frontier" essay, and says it has enhanced safeguards to combat risky behavior (Emma Roth/The Verge)

Why it matters

Launch-day availability across AWS, Devin, and agent portals gives operators more routing options for coding agents, yet competing benchmark narratives may not align with your production cost per task.

Limits and uncertainties

The Decoder notes Anthropic's parity, cost, and GPT-6 Astra benchmark claims are not independently verified in the excerpt.

Cognition's FrontierCode 1.1 ranking is a vendor benchmark and may not reflect workloads outside Devin.

The AWS ML excerpt is truncated before fully describing safety features referenced in the post.

Practical implications

AWS builders can evaluate Opus 5.5 on Bedrock for agentic coding, but should compare token pricing and cache read costs against Opus 5 using their own traces.

Devin users can select Opus 5.5 in Desktop and CLI while treating Cognition's cost-per-task claims as hypotheses until measured on real repos.

Teams on Hermes Agent via Nous Portal can trial the model, with no independent verification provided in Teknium's post beyond the announcement.

What to watch

AWS publication of full Opus 5.5 pricing and safety feature details beyond the truncated announcement excerpt.

Anthropic rollout timing for Sonnet 5.5 and Haiku 5.5 mentioned as expected in coming weeks.

Independent benchmark and cost analyses such as Artificial Analysis and Hacker News discussion threads cited in the packet.