Competitive race

GLM-4.6V open weights connect multimodal perception to tool use

Z.ai released GLM-4.6V and GLM-4.6V-Flash with MIT weights, 128K multimodal context, and native function calling that turns images, documents, and interfaces into agent actions.

CURRENT ASSESSMENT · REVISION 1
TOWARD DOOM39confidence 84/100

Why it moved the index

Open multimodal models that natively translate untrusted visual inputs into tool actions directly expand agentic diffusion and prompt-injection exposure.

AUDIT TRAIL

Assessment history

  1. R1
    Toward 39 · confidence 84

    New historical record from Z.ai's dated launch, repository, and model cards.

    11 Aug 2026