Benchmark menu

Model detail

MiMo-V2.5

A static, shareable model page that remains complete without JavaScript.

MiMo-V2.5

Provider: Xiaomi · Public slug: xiaomi-mimo-v2-5

Rank
25
Rank movement
Baseline
Score
27.7
Last measured
2026-07-05
Coverage
Complete coverage

Reasoning scores

XHigh
22.1
Medium
24.5
None
36.5

Generation health

XHigh0%63%37%n 8Medium5%67%28%n 21None50%21%29%n 24 Complete coverageHigh repair rateHigh failure rate

Cost summary

Mean standard list-price equivalent
$0.02
Median unique-cell cost
$0.01
Unique-cell range
$0.0017–$0.06
Coverage
24 unique cells / 53 observed runs
Observed-run total
$0.70
Evidence basis
Provider-reported
Mean output
41k tok
Output range
1.6k–147k tok
Token samples
53

Cost means equal-weight unique model/game/reasoning cells. Provider-account billed spend is reported separately on Charts.

Run setup

  • Overall Models: combines public reasoning variants
  • Standard GameBench code-generation prompt