Gemini 3.1 Pro
Parameters
—
Architecture
—
Released
—
License
—
About
No detailed description available for this model.
Benchmark Scores
| Benchmark | Score | Date | Standing |
|---|---|---|---|
|
AgentWorldBench MCP
general_agent
|
27.29%
|
24.06.2026 |
#10/14
better than 31%
|
|
AgentWorldBench Search
general_agent
|
52.08%
|
24.06.2026 |
#6/14
better than 62%
|
|
AgentWorldBench Terminal
general_agent
|
61.79%
|
24.06.2026 |
#9/14
better than 38%
|
|
AgentWorldBench SWE
general_agent
|
69.66%
|
24.06.2026 |
#10/14
better than 31%
|
|
AgentWorldBench Web
general_agent
|
75.79%
|
24.06.2026 |
#2/14
better than 92%
|
|
AgentWorldBench OS
general_agent
|
76.45%
|
24.06.2026 |
#4/14
better than 77%
|
|
AgentWorldBench Overall
general_agent
|
67.12%
|
24.06.2026 |
#8/14
better than 46%
|
|
SWE-bench Verified
coding_agent
|
91.97%
|
11.06.2026 |
#7/60
better than 86%
|
|
Terminal Bench 2.1
coding_agent
|
77.49%
|
11.06.2026 |
#23/44
better than 49%
|
|
SWE Atlas - QnA
coding_agent
|
7.85%
|
11.06.2026 |
#21/24
better than 13%
|
|
NL2Repo-Bench
coding_agent
|
21.60
|
11.06.2026 |
#23/23
better than 0%
|
|
SWE Atlas - TW
coding_agent
|
39.50%
|
11.06.2026 |
#13/20
better than 37%
|
|
SWE-efficiency
coding_agent
|
17.48%
|
11.06.2026 |
#4/5
better than 25%
|
|
LiveSQLBench
coding_agent
|
84.62%
|
11.06.2026 |
#4/5
better than 25%
|
|
CL-bench
coding_agent
|
57.00%
|
11.06.2026 |
#3/5
better than 50%
|
|
VIBE-V2
coding_agent
|
28.00
|
11.06.2026 |
#9/9
better than 0%
|
|
SVG-Bench
coding_agent
|
69.57%
|
11.06.2026 |
#6/9
better than 38%
|
|
PostTrainBench
coding_agent
|
7.17%
|
11.06.2026 |
#20/22
better than 10%
|
|
KernelBench Hard
coding_agent
|
40.10%
|
11.06.2026 |
#4/5
better than 25%
|
|
PaperBench
coding_agent
|
25.80%
|
11.06.2026 |
#9/10
better than 11%
|
|
GDPVal
general_agent
|
46.22%
|
11.06.2026 |
#12/20
better than 42%
|
|
BankerToolBench
general_agent
|
17.22%
|
11.06.2026 |
#14/15
better than 7%
|
|
OfficeQA Pro
general_agent
|
18.10
|
11.06.2026 |
#14/14
better than 0%
|
|
SpreadSheetBench-v1
general_agent
|
39.01%
|
11.06.2026 |
#8/9
better than 12%
|
|
YC-Bench
general_agent
|
50.00%
|
11.06.2026 |
#4/8
better than 57%
|
|
Claw-Eval Avg
coding_agent
|
69.13%
|
11.06.2026 |
#25/40
better than 38%
|
|
OSWorld-Verified
general_agent
|
82.52%
|
11.06.2026 |
#6/25
better than 79%
|
|
OmniDocBench 1.5
document_understanding
|
96.17%
|
11.06.2026 |
#10/26
better than 64%
|
|
VideoMMMU
video_understanding
|
100.00%
|
11.06.2026 |
#1/19
better than 100%
|
|
VideoMME (w sub.)
video_understanding
|
90.00%
|
11.06.2026 |
#3/17
better than 88%
|
|
IMO 2025
stem_reasoning
|
37.92%
|
11.06.2026 |
#3/4
better than 33%
|
|
Humanity's Last Exam
stem_reasoning
|
81.25%
|
17.06.2026 |
#7/71
better than 91%
|
|
CritPt (no tools)
stem_reasoning
|
53.94%
|
17.06.2026 |
#10/21
better than 55%
|
|
AIME 26
stem_reasoning
|
98.72%
|
17.06.2026 |
#6/45
better than 89%
|
|
HMMT Nov 25
stem_reasoning
|
91.37%
|
17.06.2026 |
#4/20
better than 84%
|
|
HMMT Feb 26
stem_reasoning
|
86.53%
|
17.06.2026 |
#17/44
better than 63%
|
|
IMOAnswerBench
stem_reasoning
|
82.95%
|
17.06.2026 |
#13/30
better than 59%
|
|
GPQA Diamond
stem_reasoning
|
100.00%
|
17.06.2026 |
#1/80
better than 100%
|
|
NL2Repo
coding_agent
|
35.54%
|
17.06.2026 |
#31/41
better than 25%
|
|
DeepSWE 1.1
coding_agent
|
12.76%
|
17.06.2026 |
#31/34
better than 9%
|
|
ProgramBench
coding_agent
|
52.98%
|
17.06.2026 |
#6/18
better than 65%
|
|
Terminal-Bench 2.1 (Terminus-2)
coding_agent
|
78.17%
|
17.06.2026 |
#16/44
better than 65%
|
|
Terminal-Bench 2.1 (Best Reported Harness)
coding_agent
|
11.81%
|
17.06.2026 |
#4/5
better than 25%
|
|
Tool Decathlon
general_agent
|
76.83%
|
17.06.2026 |
#4/11
better than 70%
|
|
MMMU-Pro
vision_language
|
93.76%
|
11.06.2026 |
#6/35
better than 85%
|
|
Apex-Agents
general_agent
|
43.30%
|
11.06.2026 |
#13/24
better than 48%
|
|
HLE (with tools)
stem_reasoning
|
66.84%
|
17.06.2026 |
#27/49
better than 46%
|
|
DeepSearch QA
general_agent
|
60.66%
|
23.08.2026 |
#5/9
better than 50%
|
|
Toolathlon Verified
general_agent
|
41.92%
|
23.08.2026 |
#25/35
better than 29%
|
|
MCPMark
general_agent
|
58.13%
|
23.08.2026 |
#4/15
better than 71%
|
|
Terminal-Bench 2.0
coding_agent
|
100.00%
|
23.08.2026 |
#1/23
better than 100%
|
|
SWE-bench Multilingual
coding_agent
|
85.09%
|
23.08.2026 |
#13/48
better than 74%
|
|
SciCode (subtask)
stem_reasoning
|
100.00%
|
23.08.2026 |
#1/15
better than 100%
|
|
OJBench
stem_reasoning
|
100.00%
|
23.08.2026 |
#1/22
better than 100%
|
|
LiveCodeBench v6
stem_reasoning
|
97.54%
|
23.08.2026 |
#4/49
better than 94%
|
|
CharXiv (RQ)
document_understanding
|
56.13%
|
23.08.2026 |
#12/34
better than 67%
|
|
MathVision
vision_language
|
84.56%
|
23.08.2026 |
#6/27
better than 81%
|
|
V-Star
vision_language
|
96.23%
|
23.08.2026 |
#2/20
better than 89%
|
|
SWE-Marathon
coding_agent
|
5.92%
|
17.06.2026 |
#16/17
better than 6%
|
|
MMLU-Pro
knowledge
|
100.00%
|
26.06.2026 |
#1/48
better than 100%
|
|
CodeForces
stem_reasoning
|
87.53%
|
26.06.2026 |
#4/18
better than 76%
|
|
Apex-Shortlist (no tools)
stem_reasoning
|
98.21%
|
26.06.2026 |
#2/11
better than 90%
|
|
GDPVal-AA v2
general_agent
|
71.76%
|
26.06.2026 |
#24/38
better than 38%
|
|
BrowseComp
general_agent
|
93.03%
|
11.06.2026 |
#3/45
better than 95%
|
|
USAMO 2026
stem_reasoning
|
47.58%
|
11.06.2026 |
#3/4
better than 33%
|
|
FrontierSWE
coding_agent
|
17.91%
|
17.06.2026 |
#10/12
better than 18%
|
|
AgentWorldBench Android
general_agent
|
96.36%
|
24.06.2026 |
#3/14
better than 85%
|
|
MCP-Atlas
general_agent
|
73.89%
|
11.06.2026 |
#29/54
better than 47%
|
|
SWE-bench Pro
coding_agent
|
60.42%
|
11.06.2026 |
#38/77
better than 51%
|
|
Claw-Eval Pass^3
coding_agent
|
61.54%
|
23.08.2026 |
#8/21
better than 65%
|
|
BabyVision
vision_language
|
47.85%
|
23.08.2026 |
#10/24
better than 61%
|
What is a good AgentWorldBench MCP score? What is a good AgentWorldBench Search score? What is a good AgentWorldBench Terminal score?
No comments yet
Be the first to share your thoughts!