MiniMax M2.5 vs MiniMax M2.7

Change models
IndexAll categories, out of 100.
50.2 ±7.1
51.2 ±4.5 best
GapIndex points below the top.
−1.0 · no clear gap
Top
CoverageIndex weight that has results.
80%
85% best
SpeedOutput tokens per second.
44/s best
35/s
EloLMArena rating and rank.
1359
1405 best
ReleasedMonth of release.
2026-02
2026-03
Input / 1MUS dollars per 1M input tokens.
$0.27
$0.21 best
Output / 1MUS dollars per 1M output tokens.
$1.08
$0.84 best
1K chats2K tokens in, 500 out, each.
$1.08
$0.84 best
1K documents20K tokens in, 1K out, each.
$6.48
$5.04 best
1K agent tasks200K tokens in, 20K out, each.
$75.60
$58.80 best
ContextMaximum tokens in one request.
200K best
197K
AgenticMulti-step tasks with tools.
51.6
52.0 best
CodingCode writing and repair.
50.7
52.3 best
ReasoningLogic problems and puzzles.
48.8
49.5 best
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
48.9
50.2 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
—
59.8
MathMath problems.
54.3 best
47.7

Benchmarks

13 shared
Only MiniMax M2.5: 7Only MiniMax M2.7: 40
MiniMax M2.5
MiniMax M2.7
80.0%91 / 149
84.0% best43 / 149
TaxEval v2Finance
68.2% best100 / 143
66.6%110 / 143
79.2%72 / 142
79.9% best71 / 142
82.1%60 / 135
86.6% best41 / 135
MMLU ProKnowledge
80.1%89 / 135
80.4% best88 / 135
CorpFin v2Finance
59.6%76 / 131
61.2% best55 / 131
14.9% best98 / 123
11.9%103 / 123
AIMEMath
88.8%28 / 95
91.0% best26 / 95
SWE-benchCoding
74.2% best49 / 87
73.8%51 / 87
41.6%29 / 66
47.2% best22 / 66
IOI v1Coding
6.7% best40 / 61
4.9%44 / 61
53.5%39 / 53
60.9% best23 / 53
38.6%41 / 51
48.4% best29 / 51