X · @teortaxesTex
· X / Twitter
WeirdML v2 essentially saturated These last 20% were bought with extreme progress in capabilities, I really wonder when we'll see open source doing th…
WeirdML v2 essentially saturatedThese last 20% were bought with extreme progress in capabilities, I really wonder when we'll see open source doing this. DeepSeek is still <50%Håvard Ihle: Claude Fable 5 (max) scores 91.9% on WeirdML, a new clear best score. For comparison, if we take the SOTA score on each task, we wou