benchmark
.
space
benchmarks
rankings
compare
voices
transcripts
papers
articles
head to head
GPT-5.6 Sol
vs
Gemini 3.1 Pro
11 shared benchmarks
11
wins
0
ties
0
wins
7.78%
ARC-AGI 3
0.37%
90.4%
BrowseComp
85.9%
84.6%
CharXiv
83.3%
72.7%
DeepSWE
12.0%
83.0%
FrontierMath
36.9%
94.6%
GPQA Diamond
94.3%
44.5%
Humanity's Last Exam
44.4%
83.6%
MCP-Atlas
78.2%
83.0%
MMMU Pro
80.5%
64.6%
SWE-bench Pro
54.2%
88.8%
Terminal-Bench 2.1
70.3%