Muse Ember Tops Browser Agent Benchmark at 3 Cents per Task

Browser Use's evaluation on the hardest long-horizon browser benchmark (v2.1, 180 tasks, 17 models) shows Muse Ember scoring 63.9/100, matching GPT-6 Luna medium while dominating the Pareto frontier at only about 3 cents per task.

2026-11-10 ~ 2026-11-11 · 2 related posts

• Browser use: Muse Ember matches GPT-6 Luna at 3 cents per task — WenhuChen · 2026-11-10
• Muse Ember sweeps Browser Use benchmark Pareto frontier at 3 cents per task — WenhuChen · 2026-11-11