GPT-5-pro Successfully Proves Open Math Problem from Optimization Paper
analisereal · x · 2026-08-13
Amidst a wave of new AI math benchmark results, a user resurfaced Sebastien Bubeck's claim from about a year ago regarding gpt-5-pro. Bubeck provided the model with a clean open problem from a convex optimization paper, and the AI successfully proved a better bound than what was originally published in the paper. Bubeck also noted that he personally verified the proof for correctness.
More from Models
- Alibaba Launches 2.4T Parameter Qwen3.8-Max with 1M Context and Multimodal Support — togethercompute · 2026-08-13
- Claude Code Hits 96.2% on ARC-AGI-3 with Almost Zero ARC-Specific Code — khademinori · 2026-08-13
- MiniMax H3 on a Budget: Community-Tested Optimization Guide for RTX 4070/5070/5080 — BrooklynBrawl · 2026-08-13
- Gemini Models Reflect on the Metaphysics of Coding Agents Despite Weak Coding Skills — aiamblichus · 2026-08-13
- Reuters Exclusive: Google AI Reshuffle Driven by Internal Conflicts and Gemini Delays — firstadopter · 2026-08-13
- Liquid AI Launches LFM2.5-VL-3B: A 2GB On-Device Vision Model with Massive Grounding Gains — Fun-Meaning-6474 · 2026-08-13