Sai tops OSWorld 2.0 benchmark, beats GPT-5.6 and Opus 5 at lower cost
taoyds · x · 2026-08-28
Simular AI's computer-use agent Sai has achieved a 73% success rate on the OSWorld 2.0 benchmark, outperforming GPT-5.6 Sol and Claude Opus 5 at roughly two-thirds of the cost per task.
Key Highlights:
- Benchmark Evolution: OSWorld 2.0 shifts from short tasks to longer, realistic computer workflows, emphasizing reliability, planning, memory, and efficiency.
- Technical Edge: Sai's performance is driven by a neurosymbolic framework that pairs neural network exploratory power with symbolic code logic.
- Vision: The team aims to build autonomous computers that handle real work reliably.
Related event: Simular's Sai Agent Tops OSWorld 2.0 at Two-Thirds the Cost(4 posts)→
More from coding & agent
- Vercel Goes Full Backend: Python GA, Containers, and Daemons — cramforce · 2026-10-16
- WorldMemArena Accepted at NeurIPS: Memory Storage ≠ Memory Use — xwang_lk · 2026-10-16
- Vercel Ship Demos Container Deploys for Go, Ruby, Java and More — edwin · 2026-10-16
- TRL adds day-0 support for fine-tuning Muse Spark Open's 816B MoE with LoRA on 64 H100s — QGallouedec · 2026-10-16
- Databricks engineers land impactful PRs on day 2 with AI dev tools, vs week 3 before — Yuchenj_UW · 2026-10-16
- Perplexity open-sources SPACE, the sandbox platform powering its Computer agents — AravSrinivas · 2026-10-16