OpenAI's Unreleased Astra Model Reportedly Solves 10 Major Math Problems
BorisMPower · x · 2026-08-01
OpenAI's unreleased next-generation model, Astra (speculated to be GPT-6), has reportedly made ten major advances across mathematics, quantum complexity, and theoretical computer science.
Key breakthroughs include:
- The first explicit non-sofic group
- Disproving Connes's rigidity conjecture
- Proving quantum parallel repetition for general two-player entangled games
- Proving Ehrhart’s volume conjecture
- The first improved general sphere-packing exponent since 1978
Reportedly, the core arguments were generated by Astra, which then formalized the proofs in Lean, producing machine-checkable certificates alongside a 249-page manuscript. The token cost for successful solution runs is estimated at roughly $2,000 each, signaling that scientific reasoning is rapidly becoming a genuine model capability.
Related event: OpenAI's rumored Astra model solves 10 math problems(47 posts)→
More from Models
- Gary Marcus: Pure LLMs Are Stochastic Parrots; Astra Proves Need for Neurosymbolic AI — GaryMarcus · 2026-08-02
- Speechify Simba 3.2 Tops Voice Arena as Best Real-Time AI Voice Model in Its Class — emmanuelvivier · 2026-08-02
- Google DeepMind Launches Gemini Robotics 2 for Whole-Body Robot Intelligence — emmanuelvivier · 2026-08-02
- GPT-5 Solves Hard Math Instantly, Fields Medalist Fears for Math Culture — The Decoder · 2026-08-02
- antirez Tests Domestic AI Models for Coding: DeepSeek vs GLM vs Kimi — antirez · 2026-08-01
- Users Complain About Gemini's Overactive Safety Filters Blocking Normal Chats — BigLead8814 · 2026-08-01