Op
gpt-6-astra
Benchmark & Insights
OpenAI OpenAI API
Updated Oct 4, 2026 All models
Sample size
232 runs
in window
Accuracy
79.1%
reference match · 282d
Confidence
87%
over 282 runs
Window end
Oct 4, 2026
most recent run
Input price
$10.00/MTok
prompt tokens
Output price
$50.00/MTok
completion tokens
Model insights
- 01 Among the safest of the leaders, missing only 6 of 25 "unsafe" days, and all of those were hard days most models also missed.
- 02 Balanced, with a few false alarms on "MEDIUM_SAFE" days.
- 03 At $10/$50 it is poor value: gpt-6.1-sol matches it at a fifth of the price.
Notes
Recent forecasts