Development
Hard reasoning gets a thousand times cheaper in sixteen months
Claim
Sam Altman revealed that solving a hard reasoning problem has become a thousand times cheaper in the sixteen months since o1, comparing it to GPT-5.4.
Reported metric: 1000x / 16 months
Reporting date: (not necessarily the event date)
Sources
Attribution and reporting linked by this record; links do not imply independent verification.
- Welcome to March 12, 2026 (theinnermostloop.substack.com)
Cite or correct this record
Suggest a correction — opens a GitHub draft for you to review and submit.