Navier-Stokes proof burned 130B output and trillions of input tokens, dwarfing top OpenAI users' few billion daily
scaling01 · x · 2026-09-24
Responding to a claim that 130B tokens is 'very little' for a mathematical proof, scaling01 notes that 90th-percentile heavy token users at OpenAI go through only a few billion tokens per day (input + output) — while the Navier-Stokes proof used 130B output tokens and trillions of input tokens, an enormous amount by comparison.
Related event: Navier-Stokes Proof Burned 130B Output Tokens, Far Beyond Heavy User Usage(2 posts)→
More from Fun
- OpenAI launches MentalHealthBench with 80+ clinicians; replies turn into memes — Yuchenj_UW · 2026-09-24
- Claude's over-the-top dramatic persona output goes viral again — repligate · 2026-09-24
- Claude models get personified: Haiku 'leaks' Anthropic teaming with xAI — will__ye · 2026-09-24
- Victor Taelin jokes Anthropic may not know Opus 5.5 is good — they're all using something better internally — IgorCarron · 2026-09-24
- Using AI to run Pokémon-style battles on long flights as a travel hack — TejasKumar_ · 2026-09-24
- Setting Astra's Thinking to 'Extra High' Makes It Over-Engineer Unit Tests — astralmatrix · 2026-09-24