OpenAI Reportedly Hid Funding for FrontierMath Benchmark Under NDA
BlancheMinerva · x · 2026-07-30
A commentator revealed that OpenAI funded the FrontierMath benchmark developed by EpochAI and had exclusive access to the test questions. Because OpenAI insisted on an NDA in the funding contract, this conflict of interest was not disclosed at the time, leading to severe doubts about EpochAI's credibility as an independent evaluator.
Related event: OpenAI Secretly Funded FrontierMath Benchmark(2 posts)→
More from Models
- Deep Dive: Real-world Performance and Controversies of Grok 4.5, Kimi K3 and More — eyishazyer · 2026-07-30
- Inside Grok 4.5: How Cursor Collaboration and Real Developer Data Shaped the Model — eyishazyer · 2026-07-30
- Kimi K3 Review: Praised for Fewer Refusals, but 2.8T Parameters Hinder Local Deployment — eyishazyer · 2026-07-30
- PolyAI Launches Dialog-RSN-1: Audio-Native LLM Overcoming Voice Agent Bottlenecks — matthen2 · 2026-07-30
- Maya-2-Native Tops Voice Arena's Real-Time Hindi Leaderboard — Bladerunner_7_ · 2026-07-30
- Cerebras lacks powerful open-weight models; Kimi K3 scores double on index — NielsRogge · 2026-07-30