MiMo-V2.5 vs Step 3.7 Flash

Change models
MiMo-V2.5Xiaomi
IndexAll categories, out of 100.
53.9 ±6.8
54.1 ±5.0 best
GapIndex points below the top.
−0.2 · no clear gap
Top
CoverageIndex weight that has results.
75%
85% best
SpeedOutput tokens per second.
15/s
98/s best
EloLMArena rating and rank.
1428
—
ReleasedMonth of release.
2026-04
2026-05
Input / 1MUS dollars per 1M input tokens.
$0.40
$0.20 best
Output / 1MUS dollars per 1M output tokens.
$2.00
$1.15 best
1K chats2K tokens in, 500 out, each.
$1.80
$0.975 best
1K documents20K tokens in, 1K out, each.
$10.00
$5.15 best
1K agent tasks200K tokens in, 20K out, each.
$120.00
$63.00 best
ContextMaximum tokens in one request.
1M best
262K
AgenticMulti-step tasks with tools.
54.9
57.6 best
CodingCode writing and repair.
54.9 best
54.1
ReasoningLogic problems and puzzles.
—
49.0
MultimodalTasks with images and text.
56.4
58.3 best
KnowledgeFacts and expert knowledge.
50.1 best
47.0
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
—
51.3
MathMath problems.
52.4
—

Benchmarks

3 shared
Only MiMo-V2.5: 25Only Step 3.7 Flash: 18
MiMo-V2.5
Step 3.7 Flash
29.9% best85 / 120
20.3%113 / 120
56.1%51 / 83
56.3% best49 / 83
Claw-EvalAgentic
62.3%24 / 42
67.1% best15 / 42