Large VLMs with short outputs are more efficient than small models with long sequences, and a multi-agent framework can transfer reasoning tokens from small models to approach large-model performance.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Rethinking Model Efficiency: Multi-Agent Inference with Large Models
Large VLMs with short outputs are more efficient than small models with long sequences, and a multi-agent framework can transfer reasoning tokens from small models to approach large-model performance.