On Thursday, Anthropic released Opus 4.8, the newest version of its most advanced publicly available model. The model is available at the same pricing tier as the previous Opus release.
The new model arrives just 41 days after Opus 4.7 — a notably faster upgrade cycle than Anthropic typically maintains. (The most recent Sonnet and Haiku models are three and seven months old, respectively.) The quick turnaround follows a mixed reception to Opus 4.7, which some users found disappointing, and comes amid significant new releases from competitors including OpenAI’s Codex and Google’s Gemini Flash.
Beyond benchmark improvements, Opus 4.8 places particular emphasis on how the model handles bad or uncertain data. According to Anthropic’s launch post, early testers found the new model is “more likely to flag uncertainties about its work and less likely to make unsupported claims.” A testimonial from Bridgewater Associates cited in the post noted that “Opus 4.8’s tendency to proactively flag issues with the inputs and outputs of an analysis” was the most meaningful improvement, describing it as “something other models routinely missed and left to the users to catch.”
Alongside the model, Anthropic launched Dynamic Workflows, now available in research preview. The feature is designed to help larger models like Opus manage complex tasks across hundreds of parallel subagents. As the company described it: “Claude Code alongside Opus 4.8 can now carry out codebase-scale migrations across hundreds of thousands of lines of code from kickoff to merge, with the existing test suite as its bar.”
Meanwhile, Anthropic’s more advanced Mythos model remains unreleased following cybersecurity concerns raised during a tentative preview last month. The company indicated in today’s post that this may change soon: “We’re making swift progress on developing these safeguards and expect to be able to bring Mythos-class models to all our customers in the coming weeks.”
Source: Anthropic’s Opus 4.8 Adds Dynamic Workflows for Large-Scale Agent Tasks