pith. sign in

Gym-V: A Unified Vision Environment System for Agentic Vision Research

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it
abstract

As agentic systems increasingly rely on reinforcement learning from verifiable rewards, standardized ``gym'' infrastructure has become essential for rapid iteration, reproducibility, and fair comparison. Vision agents lack such infrastructure, limiting systematic study of what drives their learning and where current models fall short. We introduce \textbf{Gym-V}, a unified platform of 179 procedurally generated visual environments across 10 domains with controllable difficulty, enabling controlled experiments that were previously infeasible across fragmented toolkits. Using it, we find that observation scaffolding is more decisive for training success than the choice of RL algorithm, with captions and game rules determining whether learning succeeds at all. Cross-domain transfer experiments further show that training on diverse task categories generalizes broadly while narrow training can cause negative transfer, with multi-turn interaction amplifying all of these effects. Gym-V is released as a convenient foundation for training environments and evaluation toolkits, aiming to accelerate future research on agentic VLMs.

fields

cs.AI 1

years

2026 1

verdicts

UNVERDICTED 1

representative citing papers

TRON: Targeted Rule-Verifiable Online Environments for Visual Reasoning RL

cs.AI · 2026-06-01 · unverdicted · novelty 6.0

TRON supplies 520 rule-verifiable online visual reasoning environments across five ability buckets that generate unlimited training instances for RL post-training, yielding consistent gains on ten external multimodal benchmarks for three vision-language models.

citing papers explorer

Showing 1 of 1 citing paper.

  • TRON: Targeted Rule-Verifiable Online Environments for Visual Reasoning RL cs.AI · 2026-06-01 · unverdicted · none · ref 26 · internal anchor

    TRON supplies 520 rule-verifiable online visual reasoning environments across five ability buckets that generate unlimited training instances for RL post-training, yielding consistent gains on ten external multimodal benchmarks for three vision-language models.