Agent Arena Long-Horizon Tests Show 55-Minute Gap Between Top Models

LMArena launched Agent Arena, benchmarking frontier models on long-horizon agent tasks spanning hundreds of conversational turns, such as porting a codebase to Rust. The results show long sessions amplify gaps between top models, with completion times differing by up to 55 minutes.

2026-11-05 ~ 2026-11-05 · 2 related posts

• Agent Arena: frontier models show 55-minute gap on long-horizon agent tasks — arena · 2026-11-05
• Agent Arena: long-horizon, hundreds-of-turns tasks expose a 55-minute gap between top models — arena · 2026-11-05