VisSearch-Bench: best multimodal agents fully solve only 26.47% of visual search tasks
antgroup · hf · 2026-11-06
Tsinghua, Ant Group, and Renmin University release VisSearch-Bench, evaluating multimodal agents on long-horizon visually intensive search tasks. The benchmark has 204 human-written tasks across 7 domains and 25 subdomains, spanning Set-wise Visual Search, Sequential Visual Search, Hybrid Wide–Deep Search, Competitive Visual Disambiguation, and Visual-Structure-Guided Search, each with reference answers, traceable evidence, and scoring rules. It focuses on visual context management—combining object appearance, details, or spatial relations with external information. Even the best systems fully solve only 26.47% of tasks.
More from Research
- Elastic's VP of AI: Treat search as test-time compute, latency is a currency to spend — gaganghotra_ · 2026-11-06
- Training-Free Speculative Decoding Cuts Latency for Full-Duplex Speech LMs — CUHKSZ · 2026-11-06
- TailSteer: Drift-Resistant Multi-Turn Image Editing from CUHK MMLab — MMLab-CUHK · 2026-11-06
- EvoMemState: State-Centric Long-Term Memory for Personal Agents — Zhishang Xiang · 2026-11-06
- Oxford Position Paper: From Autonomous Agents to Cognitive Symbionts — Oxford · 2026-11-06
- University of Michigan's 26-lecture robotics math course (ROB 501) is free on YouTube — techNmak · 2026-11-06