PyyanThe AI Index

Pyyan / Compare / Atria Dawn Preview vs DeepSeek V4 Flash Vision Exp

Atria Dawn Preview vs DeepSeek V4 Flash Vision Exp

2 of 5

Open-Weight Models · verified 16 Sept 2026

SAtria Dawn PreviewShanghai AI Laboratorycurrent
DeepSeek V4 Flash Vision ExpDeepSeekdeprecated
3 slots left
SpecificationAtria Dawn PreviewDeepSeek V4 Flash Vision Exp
Summary744B agentic weights under the MIT licence, published without a launch.Experimental, says so in the name, and shipped anyway.
Context256K128K
Input ($/Mtok)Not published$0.44
Output ($/Mtok)Not published$1.32
LicenceMITOpen weights
Max outputNot published32K
Parameters744B totalNot published
Released14 Sep 202621 Aug 2026
Model IDinternlm/Atria-Dawn-Previewdeepseek-ai/DeepSeek-V4-Flash-Vision-Exp
CategoryOpen-Weight ModelsOpen-Weight Models
OfficialShanghai AI Laboratory ↗—

Highlighted rows are where these differ.

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 →

DeepSeek V4 Flash Vision Exp

  • Retired 10 September 2026: requests now route to DeepSeek V4.1 Flash
  • Was DeepSeek's first Flash tier with image input

Best for cheap multimodal trials.

Full spec sheet →