Computer Anthology: A continuously evolving benchmark family for AI agents

Post by gramulho » Tue, Aug 04, 2026, 4:03 PM UTC

Post by abernat » Tue, Aug 04, 2026, 5:00 PM UTC

Post by dvaplima » Tue, Aug 04, 2026, 5:10 PM UTC

kmiens wrote: post #5

The comparison between harnesses is very nice. Interesting to see that using a different harness can bump the performance of the model as much as a new version (e.g., GPT 5.5+Codex ~= GPT 5.6+Terminus, at lower cost)

Post by MatheusFelipe » Tue, Aug 04, 2026, 5:14 PM UTC

Post by lorenaaraujo » Wed, Aug 05, 2026, 5:50 PM UTC