Muse Spark 1.1 scores 90.2% on Online-Mind2Web, edging past Claude Opus 4.8
DhruvBatra_ · x · 2026-07-24
Dhruv Batra says Muse Spark 1.1 was evaluated on Online-Mind2Web, a computer-use/browser-use benchmark.
- On the benchmark, Muse Spark 1.1 scored 90.2%.
- That is better than Claude Opus 4.8 (84.1%).
- It is slightly behind GPT-5.4 (92.8%), though the gap may not be statistically significant.
- Qualitatively, Muse appears to manipulate URLs more aggressively than other models:
- it uses query parameters when possible,
- searches for nested URLs directly,
- and can jump to alternative sites when blocked.
More from coding & agent
- Agent verified 'success' after double-voting: a lesson in harness vs model bugs — TejasKumar_ · 2026-10-26
- diginomica: agentic automation hype is overdone, and 'The Hackening' of enterprise security has begun — jonerp · 2026-10-26
- 'Vibe verification': ~60% of AI-written formal specs prove the wrong property — basedjensen · 2026-10-26
- Devin launches Security suite powered by Claude Mythos with $100M+ credits — DevinAI · 2026-10-26
- HackBench injects reward hacks into agent trajectories to stress-test monitors — DanielKhashabi · 2026-10-26
- Hugging Face ships hf CLI updates: deduped cache saves 19GB, faster buckets, retryable GPU jobs — yb2698 · 2026-10-26