Anthropic releases Claude Opus 4.8

Anthropic released Claude Opus 4.8, an upgrade to Opus 4.7 covering coding, agentic tool use and reasoning, priced the same as its predecessor. The gap between the two releases was 41 days, a markedly faster cycle than Anthropic’s earlier Opus updates, which the company set against competitive pressure from OpenAI’s Codex and Google’s Gemini Flash launches around the same period.

Anthropic said the model was more likely to flag uncertainty in its own work and less likely to make unsupported claims than its predecessor; the investment firm Bridgewater Associates, an early tester, said the clearest difference in practice was Opus 4.8’s tendency to proactively flag problems with the inputs and outputs of an analysis rather than proceeding silently. The release also introduced Dynamic Workflows, a research-preview feature for coordinating work across hundreds of parallel subagents, which Anthropic said could support codebase-scale migrations spanning hundreds of thousands of lines of code from planning through to a merged change.

Anthropic said work on safeguards for its more capable, unreleased Mythos-class models was nearly complete, with broader availability expected “in the coming weeks” — positioning Opus 4.8 as an interim upgrade to the generally available line rather than a step change, with the more consequential release still pending behind additional safety testing.