# VistaHop: Benchmarking Long-Horizon Visual DeepSearch

arXiv: https://arxiv.org/abs/2606.03273 · HTML: https://mlsanity.com/paper/2606.03273/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Document understanding and multimodal RAG, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 53 of its 94 references that are in the map.
- First posted on arXiv 2 Jun 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Document understanding and multimodal RAG](https://mlsanity.com/mllm/document-understanding-and-multimodal-rag/)

## Builds on (most cited first, 30 of 53)

- [BrowseComp: A Simple Yet Challenging Benchmark for Browsing Agents](https://mlsanity.com/paper/2504.12516/) (2025)
- [Can MLLMs Reason in Multimodality? EMMA: An Enhanced MultiModal ReAsoning Benchmark](https://mlsanity.com/paper/2501.05444/) (2025)
- [MMSearch-R1: Incentivizing LMMs to Search](https://mlsanity.com/paper/2506.20670/) (2025)
- [Insight-V: Exploring Long-Chain Visual Reasoning with Multimodal Large Language Models](https://mlsanity.com/paper/2411.14432/) (2024)
- [Grounded Reinforcement Learning for Visual Reasoning](https://mlsanity.com/paper/2505.23678/) (2025)
- [Perception Tokens Enhance Visual Reasoning in Multimodal Language Models](https://mlsanity.com/paper/2412.03548/) (2024)
- [CogCoM: A Visual Language Model with Chain-of-Manipulations Reasoning](https://mlsanity.com/paper/2402.04236/) (2024)
- [VLM-R3: Region Recognition, Reasoning, and Refinement for Enhanced Multimodal Chain-of-Thought](https://mlsanity.com/paper/2505.16192/) (2025)
- [Reinforcement Learning with Verifiable yet Noisy Rewards under Imperfect Verifiers](https://mlsanity.com/paper/2510.00915/) (2025)
- [VoCoT: Unleashing Visually Grounded Multi-Step Reasoning in Large Multi-Modal Models](https://mlsanity.com/paper/2405.16919/) (2024)
- [MMSearch-Plus: Benchmarking Provenance-Aware Search for Multimodal Browsing Agents](https://mlsanity.com/paper/2508.21475/) (2025)
- [TIR-Bench: A Comprehensive Benchmark for Agentic Thinking-with-Images Reasoning](https://mlsanity.com/paper/2511.01833/) (2025)
- [SenseNova-MARS: Empowering Multimodal Agentic Reasoning and Search via Reinforcement Learning](https://mlsanity.com/paper/2512.24330/) (2025)
- [MedEyes: Learning Dynamic Visual Focus for Medical Progressive Diagnosis](https://mlsanity.com/paper/2511.22018/) (2025)
- [BEARCUBS: A benchmark for computer-using web agents](https://mlsanity.com/paper/2503.07919/) (2025)
- [Promoting Efficient Reasoning with Verifiable Stepwise Reward](https://mlsanity.com/paper/2508.10293/) (2025)
- [Look Inward to Explore Outward: Learning Temperature Policy from LLM Internal States via Hierarchical RL](https://mlsanity.com/paper/2602.13035/) (2026)
- [CodeV: Code with Images for Faithful Visual Reasoning via Tool-Aware Policy Optimization](https://mlsanity.com/paper/2511.19661/) (2025)
- [AgentVista: Evaluating Multimodal Agents in Ultra-Challenging Realistic Visual Scenarios](https://mlsanity.com/paper/2602.23166/) (2026)
- [MMDeepResearch-Bench: A Benchmark for Multimodal Deep Research Agents](https://mlsanity.com/paper/2601.12346/) (2026)
- [DeepWideSearch: Benchmarking Depth and Width in Agentic Information Seeking](https://mlsanity.com/paper/2510.20168/) (2025)
- [OpenSearch-VL: An Open Recipe for Frontier Multimodal Search Agents](https://mlsanity.com/paper/2605.05185/) (2026)
- [Agent-X: Evaluating Deep Multimodal Reasoning in Vision-Centric Agentic Tasks](https://mlsanity.com/paper/2505.24876/) (2025)
- [Does Your Reasoning Model Implicitly Know When to Stop Thinking?](https://mlsanity.com/paper/2602.08354/) (2026)
- [Reasoning and Tool-use Compete in Agentic RL:From Quantifying Interference to Disentangled Tuning](https://mlsanity.com/paper/2602.00994/) (2026)
- [VisBrowse-Bench: Benchmarking Visual-Native Search for Multimodal Browsing Agents](https://mlsanity.com/paper/2603.16289/) (2026)
- [MC-Search: Evaluating and Enhancing Multimodal Agentic Search with Structured Long Reasoning Chains](https://mlsanity.com/paper/2603.00873/) (2026)
- [VTC-Bench: Evaluating Agentic Multimodal Models via Compositional Visual Tool Chaining](https://mlsanity.com/paper/2603.15030/) (2026)
- [Towards Long-horizon Agentic Multimodal Search](https://mlsanity.com/paper/2604.12890/) (2026)
- [ToolForge: A Data Synthesis Pipeline for Multi-Hop Search without Real-World APIs](https://mlsanity.com/paper/2512.16149/) (2025)

## Built on by (most cited first, 3 of 3)

- [CheckerBench: Can Long-Horizon Agents Synthesize Static-Analysis Checkers?](https://mlsanity.com/paper/2610.07557/) (2026)
- [RAG-Stress: Probing the Limits of Evidence Reliance in Retrieval-Augmented Generation](https://mlsanity.com/paper/2610.11183/) (2026)
- [Not All Error Yields to Scale: Where Scaling Stops in Vision-Language Inference](https://mlsanity.com/paper/2610.01640/) (2026)

## Abstract (arXiv)

Visual DeepSearch tasks require multimodal large language models (MLLMs) to resolve complex visual queries by repeatedly inspecting image regions, grounding reasoning in visual evidence, and connecting fine-grained clues across multiple steps. However, existing benchmarks primarily evaluate single-step visual understanding or isolated visual-query response generation. They have limited difficulty, limited search horizons, and single-pass image inspection, and thus fail to evaluate models' ability to iteratively revisit visual evidence and reason across multiple steps. In this work, we introduce VistaHop, a benchmark designed specifically to evaluate Visual DeepSearch. It evaluates repeated image inspection, visual-anchor grounding, and long-horizon evidence traversal across different visual regions. VistaHop comprises 600 images, 25 visual search scenarios, and 600 Visual DeepSearch tasks. We also propose VistaArena, a unified evaluation framework that supports tool-based interactions, including visual retrieval, image inspection, and evidence-grounded reasoning. Experiments show that even state-of-the-art MLRMs remain far from solving VistaHop, with the best-performing model, SenseNova-MARS-32B, achieving only 26.33% Pass@1. These findings highlight the importance of specialized benchmarks and improved agentic methods for Visual DeepSearch.