PyyanThe AI Index

Pyyan / Compare / Llama 4 Maverick vs MiMo-V2.6-Pro

Llama 4 Maverick vs MiMo-V2.6-Pro

2 of 5

Open-Weight Models · verified 29 Sept 2026

Llama 4 MaverickMetacurrent
XMiMo-V2.6-ProXiaomicurrent
3 slots left
SpecificationLlama 4 MaverickMiMo-V2.6-Pro
SummaryThe Western open-weight default.A trillion parameters under the MIT licence.
Context—1M
Input ($/Mtok)—Weights only
Output ($/Mtok)—Weights only
LicenceLlama licenceMIT
Max output—Not published
Parameters—1.02T total, 42B active
Released—21 Sep 2026
Model ID—XiaomiMiMo/MiMo-V2.6-Pro
CategoryOpen-Weight ModelsOpen-Weight Models
OfficialMeta ↗Xiaomi ↗

Highlighted rows are where these differ.

Llama 4 Maverick

The Western open-weight default.

Best for Western vendor.

Full spec sheet →

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 →