Periodic Labs' Neon model beats GPT-6 Astra on materials benchmarks using only 1,300 H200s
LiamFedus · x · 2026-09-16
Liam Fedus (ex-OpenAI ChatGPT lead) unveiled Neon, an open-source model mid-trained and RL'd with just 1,300 H200s plus months of experimental data, surpassing GPT-6 Astra on their analysis benchmark.
- Periodic Labs built high-throughput materials labs in Menlo Park to create a loop: labs generate fresh data, models learn from it, then help decide what to try next
- Initial focus: hard problems in materials science — superconductors, magnets, and semiconductor materials
- xiangfuml's framing: coding agents advanced fast because AI researchers are programmers who understand the task; in science, scientists hold the context, and capability-building takes many iterations between the two sides
More from AGI Musings
- Alpha School investigation: 17 campuses on probation, Bill Ackman as shadow lead, Holmes involved — benjaminjriley · 2026-11-02
- Google's chief economist: AI will transform more jobs than it kills — gains come from redesigning workflows — danielrock · 2026-11-02
- Daniel Lemire: AI is exploding demand for boring compute, server CPUs to double — lemire · 2026-11-02
- Alignment researcher: the word "aligned" implies consistencies that don't exist — davidmanheim · 2026-11-02
- By 2035, Indian Parents Will Ask How Many Agents You Manage, Not Your Salary — theteknosaur · 2026-11-02
- Under 5% of US consumers pay for AI, yet compute demand is already forcing usage cuts — NinaDSchick · 2026-11-02