Today, we’re releasing SpeedrunBench
The first benchmark that measures how fast agents can beat video games. We asked frontier models not to simply complete video games, but to speedrun them - across 10 titles like Mario Kart and Pokemon. On simpler games, agents get close
Opus 5 + higgsfield just killed every web designer on the planet.
Right now one is reading this and thinking "not me." it's you. yes. you. the one with the clean portfolio. the one with 6 years of experience. the one who thinks taste is a moat. I have bad news and worse
Today, we’re releasing SpeedrunBench
The first benchmark that measures how fast agents can beat video games. We asked frontier models not to simply complete video games, but to speedrun them - across 10 titles like Mario Kart and Pokemon. On simpler games, agents get close to
Fable 5.1 & Opus 3d sculpting _initial_ tests.
Fable 5.1 on the left, Opus 5 on the right. Custom C++ sculpting app, about 25 minutes each. A few things to tweak before I give them longer.
You have reached the end of the archive
All of Claude Opus 5