CellType's BioSimBench: 27B model beats frontier LLMs at simulating cell biology

SimplyAnnisa · x · 2026-11-04

CellType introduces BioSimBench, a benchmark testing whether LLMs can simulate biology using real lab data: predicting how cells respond to drugs, gene knockdowns, and immune signals. Even the best frontier models get only part of the response right.

CellType's own 27B model, trained in BioSimGym, outperforms frontier models across every task. Built on NVIDIA and powered by NeMo RL, framed as a major step for AI-driven biology.

Related event: CellType Releases BioSimBench: 27B RL Model Beats 12 Frontier Models at Simulating Cell Responses(7 posts)→

Original post →

More from Models

• Dev finds pi coding tool defaults to Opus 4.8 on a fresh install — lucasmeijer · 2026-11-05
• User to AI labs: skip the smarter model, make it 50% cheaper and 2x faster — haider1 · 2026-11-05
• Ripple claims $32M compute for open-weight coding SOTA; skeptics question the math — teortaxesTex · 2026-11-05
• Gemini 4 Flash (codename Hydrogen) spotted A/B testing in AI Studio and LMArena — lyraxana · 2026-11-05
• Perceptron's new 3D hand model beats ACE-Ego-Hand by 12-25% on HOT3D and ARCTIC — gui_penedo · 2026-11-05
• 0.8B open-source HaloGuard 2.0 claims streaming-guard SOTA, beating models 5-10x its size — const_reborn · 2026-11-05

Models channel →