Xiaomi
MiMo-V2.6-Pro
Xiaomi's open-weight omnimodal mixture-of-experts model (September 2026) with a 1M-token context window, released under the MIT license.
Our editorial assessment, based on the public sources listed below. Last reviewed .
- Overall rank
- #25
- Tier
- A
- Score
- 83.8
- Main category
- Overall
Strengths
- MIT-licensed open weights on Hugging Face for self-hosting
- Accepts text, image, audio and video input with a 1M-token context window
- Low API pricing via Xiaomi's MiMo platform and providers such as OpenRouter
Considerations
- Output speed is slower than average
- A 1T-parameter MoE model needs substantial hardware to self-host
- Smaller international developer ecosystem than the major US labs
Price
Open weights (MIT); API about $0.43 input / $0.87 output per 1M tokens (per Artificial Analysis)
Shown for reference — price is not part of the score.
Score breakdown
- Performance
- 88.0
- 55% of score
- Reputation
- 74.0
- 35% of score
- Recency
- 95.0
- 10% of score
The overall score combines these sub-scores using the weights shown. How we score →
Verification status
Each field carries its own re-verification date. Amber means older than 90 days — treat those figures as stale until re-checked.
- Pricing
- Benchmark
- Reputation
- Feature list
- source ↗
Checked against the sources listed below.
Sources
Similar models
- Claude Opus 4.8Anthropic#21 · Tier A · Score 84.7Compare
- Grok 4.5xAI#24 · Tier A · Score 83.8Compare
- Grok 4.6xAI#28 · Tier A · Score 83.2Compare
- Mistral Large 4Mistral AI#36 · Tier A · Score 81.8Compare