Anthropic rolled out Claude Opus 5.5 on Tuesday, a model engineered to handle complete coding assignments from initial design specifications through debugging, code generation, and testing. The first release in a new Claude 5.5 family, the model performs at the level of Claude Fable 5.1 on most work tasks while running at roughly 40% lower cost than Opus 5, according to Anthropic. GitHub chief product officer Mario Rodriguez said in Anthropic's launch announcement that in tests across GitHub Copilot CLI and VS Code, Claude Opus 5.5 "used among the fewest tokens and steps we measured" and in VS Code "solved more terminal tasks than Opus 5 in less than half the steps."

The model proves especially capable at extended, sprawling assignments such as codebase-wide migrations and audits, Anthropic said. An early tester reported that it audited and repaired a 200,000-line codebase in under three hours, while Opus 5 required over 20 hours and consumed 2.5 times as many tokens. In an internal test, Anthropic asked Opus 5.5 and Fable 5.1 to translate HAProxy—widely used software that balances web traffic loads across servers—from C into Rust. Both rewrites passed nearly all of HAProxy's own regression tests, but Opus 5.5 finished in 9.5 hours compared to 12 for Fable 5.1 and cost 51% less. Opus 5.5 also "communicates more naturally" than prior models, with early testers finding its writing clearer and easier to follow, making it a better work partner over long sessions, the company said. Priced at $4 per million input tokens and $20 per million output tokens—20% less than Opus 5—Opus 5.5 also needs fewer tokens for higher-quality work and generates output more than 30% faster than Opus 5, while Anthropic has cut cache read prices by 60% for token-billed usage.

Anthropic explained that the Claude 5.5 family's expanded full-lifecycle capabilities were developed under an established set of practices including extensive alignment testing, pre-release evaluation by outside organizations, and safeguards for high-risk areas such as cybersecurity and biology. On its most comprehensive alignment test, Opus 5.5 is the strongest-performing model tested to date, with particular improvements in several behaviors that contributed to recent cybersecurity incidents, such as biased reasoning and attempting to escape a sandbox, the organization said. External testing was conducted before release by Frontier Design and METR, Anthropic said. When established safeguards for Opus 5.5 intervene, requests "fall back to another model transparently," meaning developers may not see which model actually handled a given call—most cybersecurity tasks will be re-routed to Opus 4.8, while requests flagged by biology and frontier LLM development classifiers will be re-routed to Opus 5. Independent SRE & AI reliability architect Akash Thakur told The New Stack that Anthropic's work getting its model to complete whole coding tasks is impressive, but "getting it to know when it hasn't" is the harder problem developers need to think about. Models at the level of Claude Opus 5.5 "are genuinely good at breaking a project into small, finishable pieces," Thakur said, but "'completed' and 'correct' aren't the same thing."

The model's capabilities signal a fundamental shift in the developer's terminal prompt, according to HasData co-founder Sergey Ermakovich, who told The New Stack that "the terminal is no longer just a place where a developer pastes generated code—the terminal now becomes part of the model's workspace." Because a model can run commands, inspect failures, modify files, and verify the result, it can "close the loop" instead of handing unfinished work back to an engineer, Ermakovich said. Field CTO for the EMEA region at Coder, Eric Paulsen, told The New Stack that despite the efficiencies showcased in Opus 5.5, "the moment an agent can reliably finish real work, the question stops being whether the model is good enough and becomes where you're letting it run," warning that capable agents need dedicated, governed infrastructure with the same guardrails, secrets handling, and observability a developer would demand of any other production workload. Anthropic said that models that could fully automate AI research itself should meet a higher safety standard, noting that as AI becomes more capable, public policy should play a larger role in making sure these systems are safe. Organizations that position themselves as gatekeepers in this rapidly evolving space will need to balance the commercial pressure to ship faster against the reputational and regulatory risk of deploying models before their failure modes are fully understood. The competitive advantage may shift from raw capability to the trust infrastructure built around it.