Benchmark History
Past IQ test runs with detailed results
Jul 10, 22:51
Models: 5Cost: $66.81
Standardized IQ test suite administered via API. All models tested with identical prompt sets at temperature 0.
GPT-5.6-Sol-medium
120.0$15.3
GPT-5.6-Sol-ultra
116.7$33.1
GPT-5.6-Sol-low
105.0$9.5
GPT-5.6-Terra-medium
75.0$6.1
GPT-5.6-Luna-medium
30.0$2.8
Jul 10, 03:30
Models: 5Cost: $63.50
Afternoon run. Slight variation observed in Sol-ultra reasoning tasks.
GPT-5.6-Sol-medium
119.5$14.8
GPT-5.6-Sol-ultra
114.2$32.0
GPT-5.6-Sol-low
106.3$9.2
GPT-5.6-Terra-medium
73.5$5.5
GPT-5.6-Luna-medium
31.5$2.0
Jul 9, 21:45
Models: 5Cost: $65.20
Morning run. All systems nominal.
GPT-5.6-Sol-medium
118.8$15.0
GPT-5.6-Sol-ultra
115.5$32.5
GPT-5.6-Sol-low
104.2$9.8
GPT-5.6-Terra-medium
76.0$5.8
GPT-5.6-Luna-medium
29.0$2.1
Jul 9, 03:15
Models: 5Cost: $64.00
Afternoon run. Luna-medium showed slight improvement.
GPT-5.6-Sol-medium
119.0$14.5
GPT-5.6-Sol-ultra
117.0$33.5
GPT-5.6-Sol-low
103.5$9.0
GPT-5.6-Terra-medium
74.0$5.0
GPT-5.6-Luna-medium
32.0$2.0