benchmark
.
space
benchmarks
rankings
compare
voices
transcripts
articles
head to head
GPT-5.4
vs
Gemini 3.1 Pro
8 shared benchmarks
3
wins
0
ties
5
wins
73.3%
ARC-AGI 2
77.1%
82.7%
BrowseComp
85.9%
92.8%
GPQA Diamond
94.3%
36.24%
Humanity's Last Exam
44.4%
81.2%
MMMU Pro
80.5%
57.7%
SWE-bench Pro
54.2%
77.2%
SWE-bench Verified
80.6%
81.8%
Terminal-Bench 2.0
80.2%