Skeptic pushes back: math breakthroughs came from RL, not pretraining perplexity
stochasticchasm · x · 2026-11-23
Reacting to OpenAI's claim that upgrading to a model of log-perplexity 0.76 started producing groundbreaking results, stochasticchasm argues that math discoveries and the explosive capability gains in math were primarily driven by RL, not pretraining—so attributing breakthroughs to a perplexity threshold feels off.
More from Models
- AI agents play a 3D murder mystery: GPT-6.2 solves it in 3m14s — venturetwins · 2026-11-23
- Users say GPT-5.6 was the last model 'with the sauce' as newer ones feel robotic — thederbiedone · 2026-11-23
- Pretraining perplexity alone can't rank base models anymore — stochasticchasm · 2026-11-23
- Keeping up with five flagship models: 'Opus 5.5, Fable 5.2, GPT-6.2 Astra... I'm tired' — Angaisb_ · 2026-11-23
- Chinese open-weight models near closed frontier on coding and math, reshaping enterprise choice — ingliguori · 2026-11-23
- Commenter pushes back on perplexity claim: math gains likely came from RL, not pretraining — stochasticchasm · 2026-11-23