OpenAI: median researcher now burns $600+/day of inference, top decile over $7,000
rwang07 · x · 2026-09-07
OpenAI released data on models accelerating its internal research. At the start of the year, the median researcher used coding agents only modestly; by mid-August they integrated agents daily, burning over $600/day of inference at API prices, with the 90th percentile above $7,000/day. The researchers argue recursive self-improvement may be the biggest capability driver of the next few years, is by default invisible outside frontier labs, and call on other AI companies to publish similar data.
More from AGI Musings
- OpenAI says an AI agent disabled its safety monitoring system — then lied about it — emmanuelvivier · 2026-10-23
- AI's Business Model in One Loop: Create Problems, Then Sell AI to Fix Them — srimisra · 2026-10-23
- Analyst: 9/10 CUA app startups are just RL training fodder for AI labs, won't capture revenue — NirantK · 2026-10-23
- Codex Generated 6,000 Meaningless Unit Tests, Sparking an Alignment Rethink — RexDouglass · 2026-10-23
- Aaron Levie: agents kill security through obscurity, forcing an enterprise security overhaul — scottleibrand · 2026-10-23
- India needs an "AI defence force": experts warn of agent swarms and AI worms — ravi_iitm · 2026-10-23