Pyyan / Compare / MiMo-V2.6-Pro vs Muse Spark Open

MiMo-V2.6-Pro vs Muse Spark Open

2 of 5

Open-Weight Models · verified 20 Oct 2026

XMiMo-V2.6-ProXiaomicurrent
Muse Spark OpenMetacurrent
3 slots left
SpecificationMiMo-V2.6-ProMuse Spark Open
SummaryA trillion parameters under the MIT licence.The largest American open weights, under a licence with a revenue trigger.
Context1M1M
Input ($/Mtok)Weights onlyWeights only
Output ($/Mtok)Weights onlyWeights only
LicenceMITMuse Model License
Max outputNot publishedNot published
Parameters1.02T total, 42B active816B total, 42B active
Released21 Sep 202615 Oct 2026
Model IDXiaomiMiMo/MiMo-V2.6-Prometa-models/Muse-Spark-Open-816B-A42B
CategoryOpen-Weight ModelsOpen-Weight Models
OfficialXiaomi ↗Meta ↗

Highlighted rows are where these differ.

MiMo-V2.6-Pro

  • 1.02T total parameters, 42B active per token, a 4.1% activation ratio
  • Text, image, video and audio in, with a 1M token context
  • Shipped with 7,000+ reinforcement learning environments and the framework that trained it

Best for multimodal work on your own cluster.

Full spec sheet →

Muse Spark Open

  • Text, image, video and audio in, text out, with a 1M token context
  • Muse Model License, not open source: attribution required, and a separate agreement above $20m of model-service revenue
  • About 1.5TB in BF16, or two Mac Studios at 4-bit

Best for multimodal work you host yourself.

Full spec sheet →