Anthropic launches Opus 4.8 with new “Dynamic Workflow” tool

Anthropic on Thursday released Opus 4.8, the latest version of its model available to a more advanced audience. The model is available everywhere, with a standard price at the same level as the previous Opus edition.

The new model comes just 41 days after the release of Opus 4.7, a much faster upgrade cycle than usual for Anthropic. (The latest Sonnet and Haiku models are three and seven months old, respectively.) The quick turnaround may have something to do with the Opus 4.7’s cold reception, which some users found disappointing.

This period also saw significant new releases of OpenAI’s Codex model and Google’s Gemini Flash model, increasing pressure on Anthropic to keep up with progress.

The new version comes with best-in-class predicted benchmark results, but there is also special attention to how the model manages bad or uncertain data. In the launch post, early Anthropian testers found that Opus 4.8 is “more likely to signal skepticism about its work and less likely to make unsupported claims.”

Reiterating this point, a testimonial from Bridgewater Associates said the biggest difference in the upgrade was Opus 4.8’s “propensity to proactively report issues with analysis inputs and outputs, something other models routinely missed and left users to discover.”

Along with the new model, Anthropic has launched a new feature called Dynamic Workflows, which will be available in Search Preview. The system is designed to help larger models like Opus manage complex tasks across hundreds of parallel sub-agents.

“Claude Code combined with Opus 4.8 can now perform codebase-wide migrations across hundreds of thousands of lines of code from inception to merge, with the existing test suite as its tape,” the post explains.

Anthropic is still holding back its more advanced Mythos model after an initial preview last month raised cybersecurity concerns. However, the company hinted in today’s Opus release that the Mythos preview period may end soon, once the necessary safeguards are complete.

“We are making rapid progress in developing these safeguards and expect to be able to offer Mythos-class models to all of our customers in the coming weeks,” the company wrote.