multimodal LLM
MiMo-V2.6-Flash
Xiaomi omnimodal MoE model for long-horizon agentic coding and tool use with a one-million-token context.
Levelmodel
Classmultimodal LLM
Parameters309B
Context1,000,000
Released21 Sep 2026
Opennessopen weights permissive
Runs onnot in the notes
Statusactive
Benchmark Results
vendor-reportedindependent
Artificial Analysis Intelligence Index v4.3.2composite index
Intelligence Index v4.3.2independent
38
03060
Score in index. Higher is better.
rank 8 of 117, AA model page, rank within model class
AutomationBenchagents and tool use
AutomationBench v1.0.6vendor-reported
52.3
050100
Score in %. Higher is better.
MiMo-V2.6 model card comparison table
DeepSWE v1.1coding
vendor-reported
67.9
050100
Score in %. Higher is better.
MiMo-V2.6 model card comparison table
ProgramBenchcoding
vendor-reported
26
050100
Score in %. Higher is better.
MiMo-V2.6 model card comparison table
Terminal-Bench 2.1agents and tool use
vendor-reported
87.6
050100
Score in %. Higher is better.
MiMo-V2.6 model card comparison table
Terminal-Bench 4.0agents and tool use
vendor-reported
28.8
050100
Score in %. Higher is better.
MiMo-V2.6 model card comparison table
Toolathlonagents and tool use
Toolathlon-Verifiedvendor-reported
73.6
050100
Score in %. Higher is better.
MiMo-V2.6 model card comparison table
Where this comes from
Every score names who ran it: vendor-reported, maintainer-run or independent. A rank is shown with the population it was ranked in (gap rule 3).
R&D Topics
Runtimes
Identity
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.
Architecture and Size
- Architecture
- MoE Hybrid sliding-window and global attention backbone, 256 routed experts with 8 active, 681M-param ViT and audio encoders
- Total parameters
- 309B sources
- Active parameters
- 15B
- Checkpoint parameters
- –
- Context length
- 1,000,000 tokens
- Input
- text, image, video, audio
- Output
- text
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).
Openness and Access
- Openness
- open weights permissive
- Licence
- MIT License sources
- Access routes
- open weights download
Where this comes from
The licence file governs over registry tags and aggregator labels.
Deployment
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.
Indonesia
- Indonesian support
- not stated
- Regional languages
- not in the notes
- Jakarta hosting
- unknown
- Languages
- Hugging Face language tags: en, zh
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.
Lineage
The notes record no lineage for this model.
Industry Use Cases
No industry use case links this model.
Links
DeveloperXiaomibig tech, China
Xiaomibig tech, ChinaModelMiMo-V2.6-FlashXiaomi, multimodal LLM
MiMo-V2.6-FlashXiaomi, multimodal LLMR&D TopicsAgentic Software EngineeringAgents and Tool UseEfficiency Architectures (MoE, Sparse Attention, Long Context)Omni-Modal and Unified Multimodal Models+1 more
Use CasesProcess Optimisation and Advanced Process ControlIndustrialClinical LLM (Open Weights)HealthcareTelecom Domain LLMs and BenchmarksTelecomUse cases of models in the same R&D topic; none is recorded for this model itself.
Who Runs It
PT Amman Mineral Nusa Tenggara (AMMAN)In production, 8 Jul 2026MIND IDResearch, 14 Oct 2025
Conch CementIn production, 23 Jun 2025
PT Kalbe Morinaga IndonesiaIn production, Oct 2024+2 moreDeployments of these use cases; none names this model.
PT Amman Mineral Nusa Tenggara (AMMAN)In production, 8 Jul 2026MIND IDResearch, 14 Oct 2025Conch CementIn production, 23 Jun 2025PT Kalbe Morinaga IndonesiaIn production, Oct 2024+2 moreDeployments of these use cases; none names this model.Models 1
Organisations 1
R&D topics 5
Open-Weight Frontier Releases as Research SubstrateUsing released weights as the base for downstream research.used byEfficiency Architectures (MoE, Sparse Attention, Long Context)Reducing training and inference cost per capability.produced byAgents and Tool UseSystems that plan, call tools and act over multi-step tasks.produced byAgentic Software EngineeringTraining and evaluating models that act as agents in real repositories, reading code, run…produced byOmni-Modal and Unified Multimodal ModelsSingle models that take text, images, audio and video and produce text and speech, or tha…produced by
Sources
- huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL
- huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-MOPD https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-MOPD
- artificialanalysis.ai/models/mimo-v2-6-flash https://artificialanalysis.ai/models/mimo-v2-6-flash
To cite this page: Project AURION, AI Use-Cases Research & Innovation, "MiMo-V2.6-Flash", https://aurion.semai.id/models/xiaomi-mimo-v2-6-flash/, records of October 2026.