MiMo-V2.5 vs MiMo-V2.5-Pro

Change models
MiMo-V2.5Xiaomi
IndexAll categories, out of 100.
54.8 ±6.8
56.3 ±4.6 best
GapIndex points below the top.
−1.5 · no clear gap
Top
CoverageIndex weight that has results.
75%
85% best
SpeedOutput tokens per second.
4/s
47/s best
EloLMArena rating and rank.
1428
1464 best
ReleasedMonth of release.
2026-04
2026-04
Input / 1MUS dollars per 1M input tokens.
$0.40 best
$0.48
Output / 1MUS dollars per 1M output tokens.
$2.00
$1.80 best
1K chats2K tokens in, 500 out, each.
$1.80 best
$1.86
1K documents20K tokens in, 1K out, each.
$10.00 best
$11.40
1K agent tasks200K tokens in, 20K out, each.
$120.00 best
$132.00
ContextMaximum tokens in one request.
1M
1.02M best
AgenticMulti-step tasks with tools.
56.5
58.2 best
CodingCode writing and repair.
55.5
57.2 best
ReasoningLogic problems and puzzles.
—
53.3
MultimodalTasks with images and text.
57.2
—
KnowledgeFacts and expert knowledge.
50.5
51.2 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
—
63.7
MathMath problems.
52.7
54.8 best

Benchmarks

21 shared
Only MiMo-V2.5: 7Only MiMo-V2.5-Pro: 18
MiMo-V2.5
MiMo-V2.5-Pro
78.9% best101 / 149
77.1%111 / 149
TaxEval v2Finance
71.8%72 / 143
73.8% best46 / 143
81.5% best60 / 142
81.4%62 / 142
81.6%61 / 135
82.6% best59 / 135
MMLU ProKnowledge
82.9%75 / 135
84.6% best60 / 135
CorpFin v2Finance
59.9%71 / 131
61.4% best52 / 131
42.2% best75 / 123
34.1%79 / 123
MedScribeHealthcare
72.2%90 / 109
83.7% best42 / 109
MedCodeHealthcare
31.9%91 / 106
32.5% best88 / 106
9.1%82 / 90
15.9% best70 / 90
SWE-benchCoding
71.0%60 / 87
74.0% best50 / 87
56.1%51 / 83
57.2% best44 / 83
7.8%71 / 82
11.5% best66 / 82
36.7%67 / 79
41.5% best61 / 79
14.2%57 / 77
21.6% best50 / 77
EMBFinance
55.1%50 / 75
55.2% best49 / 75
60.7% best40 / 75
57.3%44 / 75
16.0%46 / 54
22.0% best42 / 54
46.9%26 / 52
62.7% best9 / 52
Claw-EvalAgentic
62.3%24 / 42
63.8% best19 / 42