Nov 17, 2026
OpenAI's GPT-6.2 Sol scores 66.06% on the Vals Index
We evaluated OpenAIโs new GPT-6.2 Sol on Vals Index v2.2, its components and the Vals Science Index.
-
#9 of 40 on the Vals Index (66.06%) at $3.18 per test.
-
#3 of 47 on Terminal-Bench-Science (58.57%), and #7 of 27 on the Vals Science Index (57.76%).
-
On science components, it scores 77.78% on BioMysteryBench and 36.94% on MysteryMechanism.
-
On coding and finance, it scores 60.61% on Terminal-Bench 4.0 and 74.29% on EMB v1.1.
GPT-6.2 Sol has a 1M-token context window and 128k max output tokens. We evaluated it at max reasoning effort.
Congrats to the OpenAI team on the release!