PyyanThe AI Index

Pyyan / Compare / MiMo-V2.6-Pro vs Atria Dawn Preview

MiMo-V2.6-Pro vs Atria Dawn Preview

2 of 5

Open-Weight Models · verified 29 Sept 2026

XMiMo-V2.6-ProXiaomicurrent
SAtria Dawn PreviewShanghai AI Laboratorycurrent
3 slots left
SpecificationMiMo-V2.6-ProAtria Dawn Preview
SummaryA trillion parameters under the MIT licence.744B agentic weights under the MIT licence, published without a launch.
Context1M256K
Input ($/Mtok)Weights onlyNot published
Output ($/Mtok)Weights onlyNot published
LicenceMITMIT
Max outputNot publishedNot published
Parameters1.02T total, 42B active744B total
Released21 Sep 202614 Sep 2026
Model IDXiaomiMiMo/MiMo-V2.6-Prointernlm/Atria-Dawn-Preview
CategoryOpen-Weight ModelsOpen-Weight Models
OfficialXiaomi ↗Shanghai AI Laboratory ↗

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 →

Atria Dawn Preview

  • Built on Zhipu's open GLM-5.2 rather than trained from scratch
  • 96.0 on DeepSearchQA and 86.5 on CyberGym, both reported by the lab
  • BF16 and FP8 checkpoints, served through SGLang or vLLM

Best for agentic work on your own cluster.

Full spec sheet →