Unverified: OpenAI Reportedly Trained a New Model in 1.5 Weeks That Is 4x Better Than Astra at Math
felpix_ · x · 2026-09-22
An unverified claim from user felpix: OpenAI began a new training round on August 28, started using it to solve NS four days later, and the model solved NS four days after that. Training was ongoing as of September 8. The author concludes OpenAI trained a model 4x better than Astra at math in about a week and a half, adding "we're 3 months or less away from RSI." Treat as unverified timeline speculation.
More from Models
- Hark launches Pro⁴ plan for power users with 10x the usage of Pro² — adcock_brett · 2026-11-12
- Why Is Inkling 2 Small Trained on So Few Tokens Yet Still Competitive? — georgejrjrjr · 2026-11-12
- OpenAI's five biggest failures laid out: pretraining decline, alignment fears, enterprise lost to Claude — teortaxesTex · 2026-11-12
- Fable 5.2 Hits 41.3% on Remote Labor Index, Up From 2.5% a Year Ago — alliekmiller · 2026-11-12
- Beaker automates MIT's weight-space programming: one-neuron edit lifts MATH-500 from 7% to 37% for under $4 — VictorSanh · 2026-11-12
- Researcher says Anthropic dodged questions on Mythos false positive rate; 37,966 findings, 2 exploited — OwariDa · 2026-11-12