MiniMax M2.5 vs MiniMax M2.1

Change models
IndexAll categories, out of 100.
50.2 ±7.1 best
49.1 ±8.9
GapIndex points below the top.
Top
−1.1 · no clear gap
CoverageIndex weight that has results.
80% best
65%
SpeedOutput tokens per second.
44/s best
34/s
EloLMArena rating and rank.
1359
—
ReleasedMonth of release.
2026-02
2025-12
Input / 1MUS dollars per 1M input tokens.
$0.27 best
$0.30
Output / 1MUS dollars per 1M output tokens.
$1.08 best
$1.20
1K chats2K tokens in, 500 out, each.
$1.08 best
$1.20
1K documents20K tokens in, 1K out, each.
$6.48 best
$7.20
1K agent tasks200K tokens in, 20K out, each.
$75.60 best
$84.00
ContextMaximum tokens in one request.
200K
205K best
AgenticMulti-step tasks with tools.
51.6 best
48.5
CodingCode writing and repair.
50.7
51.6 best
ReasoningLogic problems and puzzles.
48.8
—
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
48.9
51.7 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
—
—
MathMath problems.
54.3 best
43.9

Benchmarks

13 shared
Only MiniMax M2.5: 7Only MiniMax M2.1: 5
MiniMax M2.5
MiniMax M2.1
80.0% best91 / 149
75.4%115 / 149
TaxEval v2Finance
68.2% best100 / 143
66.4%112 / 143
79.2%72 / 142
81.8% best59 / 142
82.1% best60 / 135
78.5%72 / 135
MMLU ProKnowledge
80.1%89 / 135
87.0% best35 / 135
CorpFin v2Finance
59.6% best76 / 131
58.9%81 / 131
AIMEMath
88.8% best28 / 95
77.9%49 / 95
MedQAHealthcare
92.5% best30 / 95
91.2%40 / 95
SWE-benchCoding
74.2%49 / 87
74.8% best47 / 87
41.6% best29 / 66
37.1%38 / 66
IOI v1Coding
6.7% best40 / 61
2.3%56 / 61
53.5%39 / 53
55.8% best32 / 53
38.6% best41 / 51
33.4%44 / 51