OpenAI internal-task 80% time horizon is just 15 min but doubling every ~86 days
BasilHalperin · x · 2026-11-09
BasilHalperin analyzes OpenAI's released internal AI R&D task data, showing how fitting methodology changes conclusions about the 80%-success time horizon:
- Naive fit (empirical p80): 80% horizon 15 min, doubling every 585 days (rising slowly); 50% horizon 10 hours, doubling every 68 days.
- METR-style fit (logistic fit to success curves): 50% horizon 6 hours with 63-day doubling; 80% horizon 15 minutes with 86-day doubling — low reliability but rising very quickly.
Bottom line: methodology matters; under METR-style fitting, OpenAI internal time horizons double roughly every 2 months, far faster than METR's public 4-7 month estimate.
More from Models
- Artificial Analysis launches Decision Index v2 with 45% private data and calibration-based scoring — ArtificialAnlys · 2026-11-09
- GLM-4.5 Hit #1 on Hugging Face Within 48 Hours of MIT-Licensed Release — ccerrato147 · 2026-11-09
- Has any LLM release ever been decidedly worse than its predecessor? — adonis_singh · 2026-11-09
- CEOs say Google's Argon stands out for writing, calling Google the AI dark horse — danshipper · 2026-11-09
- Epoch AI adds ECI scores, projects Gemini 3 Pro time horizon at 4.9 hours — CFGeek · 2026-11-09
- Infosec researcher maps 3 low-hanging abuse paths for open-weight models — matthew_d_green · 2026-11-09