Op
gpt-5-nano Benchmark & Insights
OpenAI OpenAI API
Updated Feb 22, 2026 All models
Sample size
7 runs
in window
Accuracy
71.4%
consensus match · 7d
Confidence
65%
over 7 runs
Window end
Feb 22, 2026
most recent run
Input price
$0.05/MTok
prompt tokens
Output price
$0.40/MTok
completion tokens
Model insights
- 01 Only 7 days — undersampled, provisional.
- 02 Its calls already spread oddly toward "high" (3 of 7 days) and both misses overcall, one flipping to "unsafe" on a "safe" day; the nano tier looks like the weak end of the gpt-5 family.