MiniMax-M2.7
MiniMax
Parameters
—
Architecture
—
Released
—
License
—
About
No detailed description available for this model.
Benchmark Scores
| Benchmark | Score | Date | Standing |
|---|---|---|---|
|
Terminal Bench 2.1
coding_agent
|
56.21%
|
11.06.2026 |
#31/44
better than 30%
|
|
SWE Atlas - QnA
coding_agent
|
3.83%
|
11.06.2026 |
#23/24
better than 4%
|
|
SWE Atlas - TW
coding_agent
|
22.55%
|
11.06.2026 |
#15/20
better than 26%
|
|
SWE-efficiency
coding_agent
|
14.00
|
11.06.2026 |
#5/5
better than 0%
|
|
LiveSQLBench
coding_agent
|
33.20
|
11.06.2026 |
#5/5
better than 0%
|
|
CL-bench
coding_agent
|
15.40
|
11.06.2026 |
#5/5
better than 0%
|
|
VIBE-V2
coding_agent
|
35.48%
|
11.06.2026 |
#8/9
better than 12%
|
|
SVG-Bench
coding_agent
|
48.00
|
11.06.2026 |
#9/9
better than 0%
|
|
PostTrainBench
coding_agent
|
13.10
|
11.06.2026 |
#22/22
better than 0%
|
|
KernelBench Hard
coding_agent
|
10.50
|
11.06.2026 |
#5/5
better than 0%
|
|
PaperBench
coding_agent
|
30.60
|
11.06.2026 |
#10/10
better than 0%
|
|
DRACO
general_agent
|
6.84%
|
11.06.2026 |
#11/12
better than 9%
|
|
BankerToolBench
general_agent
|
63.90
|
11.06.2026 |
#15/15
better than 0%
|
|
SpreadSheetBench-v1
general_agent
|
91.76%
|
11.06.2026 |
#5/9
better than 38%
|
|
YC-Bench
general_agent
|
0.00
|
11.06.2026 |
#8/8
better than 0%
|
|
MCP-Atlas
general_agent
|
48.04%
|
11.06.2026 |
#46/54
better than 15%
|
|
Apex-Agents
general_agent
|
5.60
|
11.06.2026 |
#24/24
better than 0%
|
|
AgentWorldBench MCP
general_agent
|
5.87%
|
24.06.2026 |
#12/14
better than 15%
|
|
AgentWorldBench Search
general_agent
|
33.75%
|
24.06.2026 |
#11/14
better than 23%
|
|
AgentWorldBench Terminal
general_agent
|
41.62
|
24.06.2026 |
#14/14
better than 0%
|
|
AgentWorldBench SWE
general_agent
|
37.44
|
24.06.2026 |
#14/14
better than 0%
|
|
AgentWorldBench Android
general_agent
|
52.40
|
24.06.2026 |
#14/14
better than 0%
|
|
AgentWorldBench Web
general_agent
|
45.24%
|
24.06.2026 |
#9/14
better than 38%
|
|
AgentWorldBench OS
general_agent
|
10.48%
|
24.06.2026 |
#13/14
better than 8%
|
|
AgentWorldBench Overall
general_agent
|
46.12
|
24.06.2026 |
#14/14
better than 0%
|
|
Vals.ai Financial Agent 1.1 (without web search)
general_agent
|
51.30
|
04.06.2026 |
#7/7
better than 0%
|
|
Terminal-Bench 2.1 (Terminus-2)
coding_agent
|
50.89%
|
04.06.2026 |
#31/44
better than 30%
|
|
CritPt (no tools)
stem_reasoning
|
0.60
|
04.06.2026 |
#21/21
better than 0%
|
|
Vals.ai Financial Agent 1.1 (with web search)
general_agent
|
50.50
|
04.06.2026 |
#7/7
better than 0%
|
|
MMLU-Pro
knowledge
|
70.65%
|
04.06.2026 |
#26/48
better than 45%
|
|
LiveCodeBench v6
stem_reasoning
|
77.76%
|
04.06.2026 |
#29/49
better than 42%
|
|
SWE-bench Verified
coding_agent
|
85.89%
|
04.06.2026 |
#25/60
better than 59%
|
|
SWE-bench Multilingual
coding_agent
|
79.05%
|
04.06.2026 |
#23/48
better than 53%
|
|
IMOAnswerBench
stem_reasoning
|
64.91%
|
04.06.2026 |
#26/30
better than 14%
|
|
IMOAnswerBench (with tools)
stem_reasoning
|
75.10
|
04.06.2026 |
#7/7
better than 0%
|
|
ProfBench (Search)
general_agent
|
43.17%
|
04.06.2026 |
#6/7
better than 17%
|
|
OmniScience Non-Hallucination
knowledge
|
79.44%
|
04.06.2026 |
#3/10
better than 78%
|
|
Apex-Shortlist (no tools)
stem_reasoning
|
28.90
|
04.06.2026 |
#11/11
better than 0%
|
|
IFBench (prompt loose)
instruction_following
|
62.05%
|
04.06.2026 |
#10/14
better than 31%
|
|
PinchBench
general_agent
|
77.60
|
04.06.2026 |
#9/9
better than 0%
|
|
TauBench V3 Airline
general_agent
|
75.30
|
04.06.2026 |
#7/7
better than 0%
|
|
Apex-Shortlist (with tools)
stem_reasoning
|
51.90
|
04.06.2026 |
#7/7
better than 0%
|
|
Multi-Challenge
instruction_following
|
42.50
|
04.06.2026 |
#18/18
better than 0%
|
|
TauBench V3 Retail
general_agent
|
32.26%
|
04.06.2026 |
#5/7
better than 33%
|
|
TauBench V3 Telecom
general_agent
|
89.60
|
04.06.2026 |
#7/7
better than 0%
|
|
GPQA Diamond
stem_reasoning
|
85.44%
|
04.06.2026 |
#42/80
better than 47%
|
|
AA-LCR
long_context
|
83.10%
|
04.06.2026 |
#11/42
better than 76%
|
|
SciCode (subtask)
stem_reasoning
|
38.30
|
04.06.2026 |
#15/15
better than 0%
|
|
Humanity's Last Exam
stem_reasoning
|
39.77%
|
04.06.2026 |
#43/71
better than 40%
|
|
MMLU-ProX
multilingual
|
48.39%
|
04.06.2026 |
#13/19
better than 33%
|
|
TauBench V3 Average
general_agent
|
66.10
|
04.06.2026 |
#7/7
better than 0%
|
|
WMT24++ (en→xx)
multilingual
|
92.02%
|
04.06.2026 |
#7/17
better than 62%
|
|
BrowseComp
general_agent
|
57.30%
|
04.06.2026 |
#27/45
better than 41%
|
|
Claw-Eval Avg
coding_agent
|
58.97%
|
11.06.2026 |
#32/40
better than 21%
|
|
TauBench V3 Banking
general_agent
|
21.82%
|
04.06.2026 |
#14/17
better than 19%
|
|
OmniScience Accuracy
knowledge
|
33.80%
|
04.06.2026 |
#7/10
better than 33%
|
|
GDPVal
general_agent
|
25.90%
|
04.06.2026 |
#18/20
better than 11%
|
|
NL2Repo-Bench
coding_agent
|
24.95%
|
11.06.2026 |
#22/23
better than 5%
|
|
SWE-bench Pro
coding_agent
|
62.65%
|
11.06.2026 |
#31/77
better than 61%
|
What is a good Terminal Bench 2.1 score? What is a good SWE Atlas - QnA score? What is a good SWE Atlas - TW score?
No comments yet
Be the first to share your thoughts!