Z.ai’s open-weight 321B-total / 18B-active multimodal agent model. Self-host the native FP8 checkpoint on a documented multi-GPU Linux server or rented GPU Pod; use Strix for an authorized repository review.
Model size
321B total · 18B active
Format
Native FP8
Context
1,048,576 architectural maximum; practical limit depends on GPU memory
Freshness 75 out of 100. Repository age: 75 days old, Aug 25, 2026. Measures age, not quality.Freshness75/10075 days old
Popularity 92 out of 100. 6,423,926 downloads in 30 days; 3,083 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.Popularity92/1006.4M downloads / 30d
Arena Open Source Agents: rank 10 of 23. Ranked model: GLM-5.3-Flash. Export dated Nov 4, 2026. Source checked Nov 7, 2026. Human preference ranking, not hardware suitability.#10Arena · Open SourceAgents+3
Z.ai’s open-weight ~744B-total / ~40B-active reasoning model for long-running agents and coding. Its native FP8 checkpoint has a documented 8×H200 Linux vLLM path; a rented multi-GPU Pod is the practical alternative to owning that server.
Model size
~744B total · ~40B active
Format
Native FP8
Context
1,048,576 architectural maximum; practical limit depends on GPU memory
Freshness 75 out of 100. Repository age: 75 days old, Aug 25, 2026. Measures age, not quality.Freshness75/10075 days old
Popularity 86 out of 100. 2,105,648 downloads in 30 days; 2,389 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.Popularity86/1002.1M downloads / 30d
Arena Open Source Agents: rank 4 of 23. Ranked model: GLM-5.3 (Max). Export dated Nov 4, 2026. Source checked Nov 7, 2026. Human preference ranking, not hardware suitability.#4Arena · Open SourceAgents+2