PyyanThe AI Index

Pyyan / Compare / DeepSeek V4-Pro vs Atria Dawn Preview

DeepSeek V4-Pro vs Atria Dawn Preview

2 of 5

Open-Weight Models · verified 16 Sept 2026

DeepSeek V4-ProDeepSeekcurrent
SAtria Dawn PreviewShanghai AI Laboratorycurrent
3 slots left
SpecificationDeepSeek V4-ProAtria Dawn Preview
SummaryBest all-round open model of 2026.744B agentic weights under the MIT licence, published without a launch.
Context—256K
Input ($/Mtok)—Not published
Output ($/Mtok)—Not published
LicenceMITMIT
Max output—Not published
Parameters—744B total
Released—14 Sep 2026
Model ID—internlm/Atria-Dawn-Preview
CategoryOpen-Weight ModelsOpen-Weight Models
OfficialDeepSeek ↗Shanghai AI Laboratory ↗

Highlighted rows are where these differ.

DeepSeek V4-Pro

  • Tops open leaderboards on agentic coding and reasoning
  • MIT licence — no conditions
  • Pioneered aggressive cache-hit pricing at ~$0.07/M
  • DeepSeek routes all V4-Pro requests to V4.1 Flash from 14 September 2026

Best for everything open.

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 →