Skip to content

Opus 4.8 beats Mythos on agentic computer use

Illustration of a man in a lab coat labeled Opus 4.8 writing benchmark scores in a book marked OSWorld in a server room Models

OSWorld-Verified bench tests AI on real computer tasks – opening apps, filling forms, navigating browsers. Each task is pass/fail. The % is simply how many tasks the model completed successfully out of the total

83.4% vs 79.6% – a nearly 4% gap on pass/fail tasks isn't noise. It's a pattern

Public models are closing the gap faster than anyone expected. If Anthropic doesn't open Mythos to the mass market soon, the window closes

And according to the Opus 4.8 release notes – it already is. Mythos goes public in weeks

Benchmark table comparing Opus 4.8 and Mythos on agentic coding, multidisciplinary reasoning, and agentic computer use scores
Portrait of Nick Trenkler, author at thehype.news

Nick Trenkler

Contributing Journalist & Content Creator at thehype. Nick holds a Bachelor’s degree in Journalism, Master’s in Marketing. He has been leading marketing for AI startups since 2024, with experience in media, digital strategy, and tech-focused content.

More articles by Nick Trenkler →
On air · radio.thehype.news ↗ AI news radio — 24/7