OpenAI Data Shows Reasoning Scaling Still Logarithmic

Oxford scholar and DeepMind researcher Toby Ord notes that a scaling chart in OpenAI's Navier-Stokes post shows reasoning performance still grows only logarithmically, with diminishing returns even at massive agent-scale compute.

2026-09-14 ~ 2026-09-14 · 2 related posts

• OpenAI's 10,000-agent inference-scaling graph still shows logarithmic gains, says Toby Ord — tobyordoxford · 2026-09-14
• Toby Ord flags weak inference scaling: math gains still logarithmic at 10,000 agents — tobyordoxford · 2026-09-14