DeepSeek V3.2 Thinking vs MiniMax M2.1

Change models
IndexAll categories, out of 100.
49.0 ±8.9 best
48.8 ±8.9
GapIndex points below the top.
Top
−0.2 · no clear gap
CoverageIndex weight that has results.
65%
65%
SpeedOutput tokens per second.
—
34/s
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
—
2025-12
Input / 1MUS dollars per 1M input tokens.
—
$0.30
Output / 1MUS dollars per 1M output tokens.
—
$1.20
1K chats2K tokens in, 500 out, each.
—
$1.20
1K documents20K tokens in, 1K out, each.
—
$7.20
1K agent tasks200K tokens in, 20K out, each.
—
$84.00
ContextMaximum tokens in one request.
—
205K
AgenticMulti-step tasks with tools.
47.4
48.2 best
CodingCode writing and repair.
48.8
51.1 best
ReasoningLogic problems and puzzles.
—
—
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
50.8
51.3 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
—
—
MathMath problems.
51.9 best
43.4

Benchmarks

15 shared
Only DeepSeek V3.2 Thinking: 1Only MiniMax M2.1: 3
DeepSeek V3.2 Thinking
MiniMax M2.1
76.1% best112 / 149
75.4%115 / 149
TaxEval v2Finance
68.2% best100 / 143
66.4%112 / 143
80.7%65 / 142
81.8% best59 / 142
80.3% best66 / 135
78.5%72 / 135
MMLU ProKnowledge
84.9%59 / 135
87.0% best35 / 135
CorpFin v2Finance
51.0%105 / 131
58.9% best81 / 131
AIMEMath
84.6% best39 / 95
77.9%49 / 95
MedQAHealthcare
93.9% best20 / 95
91.2%40 / 95
SWE-benchCoding
67.6%69 / 87
74.8% best47 / 87
MGSMMultilingual
86.0%65 / 74
87.9% best56 / 74
36.0%41 / 66
37.1% best38 / 66
IOI v1Coding
10.7% best34 / 61
2.3%56 / 61
55.4%34 / 53
55.8% best32 / 53
2.3%49 / 51
33.4% best44 / 51
40.0%22 / 47
41.3% best18 / 47