Anthropic's Opus 4.8 launches with Dynamic Workflows tool, faster cycle

Anthropic released Opus 4.8 on Thursday, the newest version of its most advanced publicly available model, available everywhere at standard pricing matching the previous Opus release. The update came just 41 days after Opus 4.7, a much faster upgrade cycle than normal for Anthropic (the last Sonnet and Haiku updates are three and seven months old). The rapid turnaround likely responds to a chilly reception for Opus 4.7, which some users found disappointing, and increased competition from OpenAI’s Codex and Google’s Gemini Flash. Opus 4.8 delivers expected best-in-class benchmark results, with particular attention to managing bad or uncertain data. Early testers report the model is “more likely to flag uncertainties about its work and less likely to make unsupported claims.” A testimonial from Bridgewater Associates noted that the biggest difference was “Opus 4.8’s tendency to proactively flag issues with inputs and outputs of an analysis, something other models routinely missed and left to users to catch.”

Alongside the model, Anthropic launched a feature called Dynamic Workflows, available in research preview. The system helps larger models like Opus manage complex tasks across hundreds of parallel subagents, enabling codebase-scale migrations across hundreds of thousands of lines of code from kickoff to merge, using the existing test suite as the bar. Meanwhile, Anthropic is still holding back its most advanced Mythos model after a tentative preview last month raised cybersecurity concerns. However, the company hinted in today’s release that the Mythos preview period might soon end: “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.”