An agentic MoE model for research, coding, tool use, and cybersecurity tasks.
A 744B MoE foundation model built for multi-step task completion and verifiable results.
Atria-Dawn-Preview is a 744B-parameter MoE GLM-5.2 foundation model designed for research and engineering workflows. It handles environmental understanding, tool use, and multi-step task completion across discovery, creation, delivery, and cybersecurity.
The model drives open-ended problems toward verifiable results by combining task objectives with environmental feedback. It supports the full loop of analysis, solution design, tool use, code implementation, and experiment execution with a 256K context window.
It suits researchers and engineers building autonomous agents for complex problem-solving and software development. It does not suit casual users or those lacking the hardware resources required to run a massive 744B MoE model locally.
Retrieves and organizes evidence, conducts deep research, and turns questions into executable experimental plans.
Builds software, interactive applications, games, data visualizations, and machine learning systems.
Analyzes security issues, validates vulnerabilities, applies fixes, and performs re-validation in authorized environments.
Includes specific configuration patterns for Codex and Claude Code to handle text-only model constraints.
The model is free under its MIT license, plus any API usage costs from hosted providers or the hardware infrastructure required to run it yourself.
880 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
These reviewed AI tools do the same job instantly — from our scored directory.
It is built on a 744B-parameter MoE GLM-5.2 foundation model developed by the Shanghai Artificial Intelligence Laboratory.
No, Atria-Dawn-Preview is a text-only model and requires client-side restriction settings when used with coding agents that expect multimodal input.
See Zekai first in Google