OpenAI's MedChemBench shows GPT-6.1 models near-perfect on medicinal chemistry reasoning
gdb · x · 2026-11-21
OpenAI released MedChemBench, a benchmark measuring frontier models' reasoning for identifying promising drug candidates. Greg Brockman called the progress in medicinal chemistry "very strong".
The shared results chart (average credit % vs average output tokens) shows:
- GPT-6.1 Sol/Terra and GPT-6 Astra lead, hitting 90–95%+ within a few thousand tokens;
- GPT-6 Sol lands mid-pack around 73%;
- GPT-5.6 Sol reaches 45–50%;
- GPT-5.6 Luna/Terra and GPT-6 Luna stall near 20% even at 60k output tokens.
The generational gap on molecular structure tasks is stark — newer models score far higher with far fewer output tokens.
Related event: OpenAI Launches MedChemBench for AI Drug Chemistry Reasoning(4 posts)→
More from Models
- Thinking Machines' 26B looped-MoE Inklings 2 Mini beats GLM-5.2 and Opus 4.5 on benchmarks — ziqiao_ma · 2026-11-21
- NDIF hosts Thinking Machines' Inkling 2 Mini and launches a Looped Lens interpretability challenge — benno_krojer · 2026-11-21
- Mira Murati's Thinking Machines releases Inkling 2 Mini, a looped transformer that runs on a laptop — simonguozirui · 2026-11-21
- 'Time to Announce Math 3': AI Community Memes After Math Proof Breakthroughs — burny_tech · 2026-11-21
- Agent Arena Index debuts: Gemini 4 Neon lands #2 with 1336 points — arena · 2026-11-21
- Waddle-1 robot-use agent scores 81 on unseen bimanual tasks, beating GPT-6.2 Astra at 66 — ycombinator · 2026-11-21