Ethan Mollick: OpenAI wasn't blindsided by mixed GPT 6.2 tester reactions
emollick · x · 2026-11-19
Ethan Mollick pushes back on the GPT 6.2 (Astra) discourse: 6.2 is still the best model for many use cases, model diversity in strengths is healthy, and a new Astra version plus tweaks are coming soon. He notes OpenAI encourages candid early-tester feedback and wasn't blindsided by mixed reactions, advising people not to overindex on an unusually weird point release.
Related event: Mollick Responds to GPT 6.2 Backlash(2 posts)→
More from Models
- Long thread: OpenAI lost by over-optimizing sparse terminal rewards, Astra 6.2 as proof — ivan_bezdomny · 2026-11-19
- antirez: LLM benchmarking is broken—we need jointly lab-funded trusted third parties — antirez · 2026-11-19
- Stanford releases all 57 intermediate checkpoints of Marin 535B-A23B, 18.2T tokens, 61TB of weights — stanfordnlp · 2026-11-19
- SolidWorks-Bench: no AI model can assemble a single machine; Fable 5.2 leads at 57.3% — DevvMandal · 2026-11-19
- Kimi K3.1 Pro tops open-weight models at 77% on LongTail-SWE; DeepSeek V4.1 Flash hits 76% — Fabulous_Pollution10 · 2026-11-19
- Federal audit: 4 major US AI models produced info that could help plan mass-casualty attacks — Polymarket · 2026-11-19