Leaked chart shows OpenAI internal model scoring 0.875 on BenchCAD 2.0 vs GPT-6 Astra's 0.594
ChrisGPT · x · 2026-11-04
A circulating chart attributed to OpenAI shows an unnamed internal model hitting 0.875 mean score on BenchCAD 2.0, versus 0.594 for the released GPT-6 Astra — a 28.1-point jump, or roughly 47% relative improvement. The chart also plots score against a relative cost proxy (log scale), with the internal model clearly breaking away from Astra's scaling curve at higher cost tiers. The poster speculates this is a second glimpse of GPT-6.2's benchmark performance. Unverified: this is third-party speculation, not an official OpenAI announcement.
More from Models
- Mystery Stealth Model Crouton Alpha Free for a Week With 1M Context; Guessed to Be Meta's Wonder Successor — imjustnewatai · 2026-11-04
- Particle-1 lands in OpenClaw with ~45% fewer output tokens than DeepSeek V4.1 Flash — garrytan · 2026-11-04
- Kimi K3.1 Pro revealed: 2.8T-param MoE with 1M context, weights drop tomorrow — togethercompute · 2026-11-04
- iOS 27.3 Beta surfaces in developer channel with hints of App Intents changes — rudrank · 2026-11-04
- Frontier LLM safeguards block real work; an engineer is building a fix — julianweisser · 2026-11-04
- The 'Elon Stack' infographic maps 12 Musk services taking on ChatGPT, WhatsApp, PayPal and Uber — nima_owji · 2026-11-04