CS1: a 4B clinical decision model trained on 12.4M questions beats open rivals
AI4Pathology · x · 2026-10-21
The AI4Pathology team announced Clinical System One (CS1), a fast, general-purpose decision model for clinical records, with a public preprint.
Health systems hold decades of notes, reports, and timelines containing answers rarely asked at scale—trial eligibility, missed follow-ups, drug outcomes—too slow and expensive to read even with LLMs. CS1 is a small, fast decision model tailored to clinical records:
- Trained on 12.4 million questions across 3.6 million patients
- CS1-4B beats open decision models on reading real-world patient records across diverse health systems
- An accompanying video demonstrates its use
Related event: AI4Pathology Releases CS1, a 4B Clinical Decision Model(3 posts)→
More from Research
- Spherical Fixed-Point Reasoners: looped transformers that know when to stop — LucaAmb · 2026-10-22
- OPAL: agent-in-the-loop reactive optimizers may end fixed learning-rate schedules, says Sitzmann — vincesitzmann · 2026-10-22
- First-token cueing probe shows many LLM factual errors are recall failures, not missing knowledge — SeoulNatlUniv · 2026-10-22
- OPAL: a coding agent rewrites your optimizer mid-training, live — vincesitzmann · 2026-10-22
- OPAL auto-designs DeepSeek-grade learning rate schedules for anyone — vincesitzmann · 2026-10-22
- OPAL in action: agent resets momentum and rolls back after overshooting — vincesitzmann · 2026-10-22