BenchLeader
XiaomiOpen weights: XiaomiMiMo/MiMo-V2.5Reasoning modelAuto-detected

MiMo-V2.5

Best configuration ranks #125 of 610 on the BenchLeader Index at 58.3 ±3.4. Released 22 Apr 2026.

Blended price
$0.175/M
$0.140 in · $0.280 out
Output speed
54 tok/s
First answer
45 s
Context
1M
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index58
  2. Agents & tools70
  3. Knowledge59
  4. Instruction following66
  5. Multimodal60
  6. Long context63
  7. Composite57

Benchmark results

One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

BenchmarkdefaultSource
GPQA Diamond (AA)not in index85.0%#129Artificial Analysis
Humanity's Last Exam (AA)not in index27.2%#128Artificial Analysis

Coding

BenchmarkdefaultSource
SciCode (AA)not in index43.9%#97Artificial Analysis

Agents & tools

Knowledge

BenchmarkdefaultSource
AA-Omniscience-9.8#138Artificial Analysis

Instruction following

BenchmarkdefaultSource
IFBench67.1%#88Artificial Analysis

Multimodal

BenchmarkdefaultSource
MMMU-Pro75.4%#71Artificial Analysis

Long context

BenchmarkdefaultSource
AA-LCR73.0%#132Artificial Analysis

Composite

BenchmarkdefaultSource
AA Intelligence Index22.3#148Artificial Analysis

Price history

Listed price per 1M tokens over time, as recorded by OpenRouter for the provider with the longest history.

$0.00$0.440$0.880$1.32$1.76Jul 26Aug 26
input outputnow $0.119 in · $0.238 out

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.

WorkloadTokens in / outCostWith cachingTime
Chat reply400 / 300$0.000151.0 s
Summarise a 30-page report12,000 / 600$0.001856.5 s
Code edit6,000 / 1,500$0.00131.2 min
Agentic coding session60,000 / 4,000$0.00952.0 min
Structured extraction2,000 / 200$0.000349.1 s

See also

Data as of 9 Sept 2026. Compare with another model.