Simular's Sai Agent Tops OSWorld 2.0 at Two-Thirds the Cost

Simular AI's computer-use agent Sai topped OSWorld 2.0 with 73% success, beating GPT-5.6 Sol (62.57%) and Claude Opus 5 at only two-thirds of their cost. Sai uses a neurosymbolic architecture.

2026-08-28 ~ 2026-08-28 · 4 related posts

• Sai Agent Tops OSWorld 2.0 Benchmark at Lower Cost — TianbaoX · 2026-08-28
• Simular's Agent Sai Tops OSWorld 2.0, Beating GPT-5.6 at Two-Thirds the Cost — xwang_lk · 2026-08-28
• Simular's Neurosymbolic Agent Tops OSWorld 2.0 Benchmark — xwang_lk · 2026-08-28
• Sai tops OSWorld 2.0 benchmark, beats GPT-5.6 and Opus 5 at lower cost — taoyds · 2026-08-28