← All models

Xiaomi: MiMo-V2.5-Pro

Xiaomi

balanced

Capabilities

ToolsStructured OutputExtended ThinkingLong ContextCode

Strengths

  • Exceptional long-context performance (1M+ tokens)
  • Strong agentic and tool-use capabilities
  • Excellent code generation and software engineering
  • Highly affordable for a flagship-class model

Weaknesses

  • Very slow inference (68 tok/s) limits real-time applications
  • Closed training data limits reproducibility
  • Limited provider sustainability transparency
  • No multimodal vision capabilities

Pricing

Input / 1M tokens

$0.435

Output / 1M tokens

$0.87

Context window

1,050k

Transparency

Open weights

10.0 / 10

Open training data

0.0 / 10

Open methodology

6.5 / 10

Licence openness

7.0 / 10

Provider disclosure

5.5 / 10

FMTI company score

N/A

Composite:6.0 / 10

Open-weight model from Xiaomi with moderate methodology disclosure. Training data sources not fully published. Weights publicly available; can be self-hosted for full data control.

Sustainability

Inference energy

N/A

Training footprint

N/A

Provider infrastructure

5.0 / 10

Composite:5.0 / 10

Xiaomi's infrastructure sustainability commitments not well-documented. Model size and inference costs unknown; self-hosting enables local control.

Task Fitness

question answering
79%
code
81%
mathematics
88%
business communication
86%
vision
0%
analysis
79%
extraction
91%
long-form generation
88%
research
84%
reasoning
79%
summarisation
85%
translation
82%
conversation
90%