Newest papers

OpenVLHarness: Real World Visual Agentic Tool-Use with Multimodal Memory

Yu Zhou, Cheng-Fu Yang, Rui Sun, Di Wu, Sibo Peng, Mingyang Zhang, Zhen Yang, Ruiyi Zhang, Nanyun Peng, Zhe Gan, Kai‐Wei Chang · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26860 · cited by 0 works (OpenAlex) · open-access PDF

Portuguese

Os modelos multimodais de dados de grande linguagem (MLLMs) fizeram progressos rápidos no raciocínio visual, mas muitas tarefas exigem evidências que são difíceis de obter através de raciocínio interno sozinho: nosso estudo controlado em 22 benchmarks mostra que simplesmente aumentando o orçamento de inferência do modelo de base produz ganhos modestos. Essas tarefas exigem percepção baseada, conhecimento externo e medições numéricas obtidas por meio de ferramentas heterogêneas, cujas saídas devem ser tornadas interpretáveis ao modelo e retidas em etapas de raciocínio aberto. Introduzimos o OpenVLHarness, um recurso de raciocínio multimodal organizado em torno de uma camada de capacidade e de memória multimodal persistente. A camada de capacidade combina

Machine translation (Meta M2M100); the English original follows.

Abstract

Multimodal large language models (MLLMs) have made rapid progress on visual reasoning, yet many tasks require evidence that is difficult to obtain through internal reasoning alone: our controlled study on 22 benchmarks shows that simply increasing the inference budget of the base model yields only modest gains. Such tasks demand grounded perception, external knowledge, and numerical measurements obtained through heterogeneous tools, whose outputs must be made interpretable to the model and retained across reasoning steps. We introduce OpenVLHarness, a multimodal reasoning harness organized around a capability layer and persistent multimodal memory. The capability layer combines task-level tool interfaces with evidence rendering, providing consistent access to heterogeneous backends and converting their outputs into model-readable observations. Persistent multimodal memory retains the corresponding images, structured data, and web references as addressable artifacts throughout each query. A coding interface executes generated programs over stored data, enabling computations that combine results from multiple tools. Together, these mechanisms support an iterative process of acquiring, inspecting, and composing multimodal evidence. Evaluated on 22 benchmarks with eight open and proprietary backbones, OpenVLHarness improves average scores by 9.7-12.6 points over matched no-tool baselines, including +11.2 with Qwen3-VL-8B, +9.7 with Kimi K3, and +9.9 with GPT-6 Astra. With GPT-6 Luna at medium reasoning effort, OpenVLHarness outperforms the base model, Codex, and Visual Sketchpad at maximum effort by 7.7-8.5 points while incurring at most 37% of their model-token cost.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Coding (social sciences), Inference, Visual reasoning, Multimodality, Process (computing)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange