Glance reads structured visual answers from a frozen 4B VLM, cutting GPU cost up to 85%

multiply_matrix · x · 2026-09-22

Yohei Nakajima released a working paper and tool, glance-vlm: instead of generating text, it reads structured visual judgements (yes/no, pick-one, ratings) straight from the logits of a frozen Qwen3-VL-4B.

All experiments pre-registered, results regenerate from the repo; install with pip install glance-vlm.

Original post →

More from Research

Research channel →