Development
A cheap model nearly matches the frontier at a twentieth the cost
Claim
On CursorBench 3.1 Opus 4.7 Max edged ahead at 64.8% while Cursor's own Composer 2.5 nearly matched it for a twentieth of the cost, and Fable 5 aced a physics-demo gauntlet but billed six times more than Opus 4.8.
Reported metric: 64.8% / 20x cheaper
Reporting date: (not necessarily the event date)
Sources
Attribution and reporting linked by this record; links do not imply independent verification.
- Welcome to July 2, 2026 (theinnermostloop.substack.com)
Cite or correct this record
Suggest a correction — opens a GitHub draft for you to review and submit.