Lab claims Gemini 4 Hydrogen beats Opus 5.5 on most rare-disease benchmarks at fraction of price
danielmckinn0n · x · 2026-11-13
Developer @danielmckinn0n reports (unverified) that Google DeepMind's Gemini 4 Hydrogen outperforms Anthropic Opus 5.5 on most of his lab's rare-disease benchmarks at lower cost, latency, and higher throughput:
- RareBench: 49.2% vs 51.6% (slightly behind)
- LabBench: 46.0% vs 40.2%
- LabRobotBench: 31.7% vs 23.0%
- ScreenBench: 70.2% vs 47.2%
He highlights that the speed-plus-intelligence combo lets Hydrogen act as a real-time agent in the loop with lab robotics, showing a workcell autonomously running an experiment from a single natural-language prompt.
More from Embodied
- BoSS benchmark released: 180K+ stellarators, 3TB of simulation data for generative models — ZhongingAlong · 2026-11-13
- Fly brain hooked up to a synth learns to sing Bad Apple — Bedrovelsen · 2026-11-13
- CoRL takeaways: SuperMimic steals the show, dexterity and tactile sensing improve, human data still a bottleneck — KyleMorgenstein · 2026-11-13
- Outworld Labs launches simulation platform with new physics engine for robot learning — ycombinator · 2026-11-13
- YC-backed Mantle launches AI-native maintenance service for robots with 250+ technicians — ycombinator · 2026-11-13
- UMI diffusion policy hits 100/100 cube-stacking success with slowed demos — DominiqueCAPaul · 2026-11-13