ZwZ2-27B

inclusionai/zwz2-27b

A visual-reasoning model in the Zooming without Zooming v2 family, trained from Qwen3.6-27B to connect visual evidence across image regions without inference-time zooming tools.

Compare

Context window: -
Input: Image · Text
Output: Text
Released: —

Weights: Open
Capabilities: Open weight, Reasoning

Pricing

No official pricing available.

Performance

Overall capability

25.0/100

Coding: —
Reasoning: —
Professional work: —
Knowledge & language: —

No evaluation results available.

Providers

No providers found

ModelRegistry · Catalog updated 2026-10-30