Aleph Alpha's open-weight English-German mixture-of-experts model for self-hosting, released under Apache 2.0 in October 2026
MiMo-V2.6-Pro
New· 4Xiaomi's open-weight flagship AI model with a 1-million-token context and multimodal input
About MiMo-V2.6-Pro
Xiaomi announced MiMo-V2.6-Pro on 22 September 2026 as the top model of its MiMo-V2.6 series, and released the weights openly under the MIT licence. For teams that want to run a very large model themselves, or simply inspect it, that licence is the headline: the checkpoint can be used and modified with few strings attached.
A sparse design with 42 billion active parameters
MiMo-V2.6-Pro is a sparse mixture-of-experts model. It holds 1.02 trillion parameters in total, but only 42 billion of them are active for any given token, which is how a model of this size stays practical to serve. It reads a context of up to one million tokens and can write up to 128,000 tokens in a reply. Input can be text, images, video or audio, and the output is text.
Open weights and the hosted API are not guaranteed to match
The weights appeared on the model host on 21 September under the name MiMo-V2.6-Pro-RL. The model card says the model is also available through Xiaomi's API but does not say whether the open checkpoint is identical to the one the API serves, so anyone who needs the two to behave the same should test both.
What Xiaomi offers around it
Xiaomi's overseas price list sits the model in the same row as the older mimo-v2.5-pro, which is scheduled for deprecation on 21 October 2026 at 10:00 Beijing time. Xiaomi recommends that existing users switch as soon as they can. The list also carries cached-input rates, a batch interface billed at half the normal token rates, and a separately billed UltraSpeed variant that Xiaomi describes as up to 20 times faster. That speed claim is Xiaomi's own and has not been measured independently.
Independent and vendor-reported results
Artificial Analysis, measuring the model independently, scored it 46 on version 4.3.2 of its Intelligence Index, which put it at the top of the open-weight models listed there (114 of them on 22 September), with about 124.5 output tokens per second measured. Xiaomi also publishes its own results on DeepSWE, Terminal-Bench and OSWorld-Verified; these come from Xiaomi's own evaluation and should be treated as vendor figures rather than verified scores.
TrustList summary
MiMo-V2.6-Pro is the flagship model in Xiaomi's MiMo-V2.6 line, and its weights are released under the MIT licence. The architecture is a sparse mixture-of-experts design with 1.02 trillion total parameters and 42 billion active parameters. It has a 1-million-token context window, accepts text, images, video and audio as input, and returns text of up to 128,000 output tokens. The model was announced on 22 September.
Main features
- Text generation: returns text output of up to 128,000 tokens.
- Multimodal input: accepts text, images, video and audio.
- Long context: a 1-million-token context window.
- Open weights: released under the MIT licence, though the model card does not say whether the open checkpoint matches the model the API serves.
Key facts
- Made by: Xiaomi
- Announced: 22 September
- Licence: MIT for the open weights
- Architecture: sparse mixture of experts
Summary written by TrustList from the business's own published information, Nov 2026.
Benchmarks & AI stats
| Benchmark | Official | Community avg |
|---|---|---|
| Artificial Analysis Intelligence Index v4.3.2llm | 46 index | - |
| Artificial Analysis Intelligence Index v5.0.1llm | 30 index | - |
“Official” values are editor-approved and feed the ranking. “Community avg” is the mean of member submissions (shown for transparency; it never affects the ranking until an editor approves a value).
Sign in to add a benchmark score for this model.
MiMo-V2.6-Pro launches
Trust Score
An earned signal from verification, reviews, awards, transparency and engagement - the vendor can't buy it.
Updated 9/22/2026
Request a demo or quote from MiMo-V2.6-Pro
Reviews
Write the first review of MiMo-V2.6-Pro
Used it? Your experience helps other buyers decide.
Write a reviewQuestions & answers
No questions yet. Be the first to ask about MiMo-V2.6-Pro.
MiMo-V2.6-Pro alternatives
Similar listings buyers compare against this one.
Google's open multimodal embedding model that maps text, images, audio and video into one vector space, under Apache 2.0
Open-weight translation model from Bilibili's Index team, covering 150 languages and following glossary and style instructions
Microsoft AI's text-to-speech model for keeping one voice across 23 languages, with voice cloning from short audio
Microsoft AI's real-time speech-to-text model for voice agents, transcribing 60 languages with continuous language detection
Liquid AI's open-weight decision model, giving yes or no, choice or score answers from text and images
Open-source 2B decision model from AWS's Strands Agents project that picks options and returns scores with a reliability figure
Cantina Security's first open-weight security model, a post-trained GLM-5.3-Flash for vulnerability research and exploit checks
More on TrustList
Everything here links back to the same verified catalogue. Pick your next stop.
- Write a reviewHelp other buyers decide
- CompaniesAgencies, consultancies and IT service providers, ranked by verified reviews.
- ProductsSoftware and SaaS with pricing, features, integrations and alternatives.
- ArticlesGuides, comparisons and research from the editorial desk and the community.
- AwardsAnnual recognition decided by verified reviews and an independent jury.
- LaunchesNew products and releases, voted up by the community every day.
- RequestsBuyers describe what they need; vendors respond directly.
- PeopleReviewers, authors and makers with public profiles.
- ComparePut up to four listings side by side before you shortlist.