IMBench: A Benchmark for Intuitive Robotic Manipulation

2026-07-17

AI Summary

Existing benchmarks struggle to evaluate the integrated capabilities of robots across perception, physical reasoning, and action execution, and fail to quantify human-like intuitive manipulation. This work proposes the first evaluation framework that formalizes intuitive manipulation as a cross-modal integration capability, introducing a benchmark comprising 35 tasks and 14K curated trajectories. The benchmark explicitly incorporates contact-rich manipulation, tool use, and multi-stage dependencies, accompanied by an extensible procedural scene generation toolkit. Experiments reveal that current vision-language models lack executable planning abilities, while state-of-the-art vision-language-action models still fall short in satisfying physical constraints and generalizing effectively—highlighting intuitive manipulation as a critical gap in general-purpose robotic policies.

Abstract

Humans combine reasoning and motor control to solve complex manipulation tasks under diverse constraints. They build an understanding of the physical world that helps them convert reasoning into actions and quickly adapt to new scenes, tasks, and rules. We refer to this capability as intuitive manipulation. Existing benchmarks fail to capture this integration: they evaluate physical reasoning in isolation from execution, or measure policy performance without requiring explicit reasoning. We introduce IMBENCH, a benchmark designed to evaluate intuitive manipulation as an integrated capability spanning perception, physical reasoning, action generation, and iterative execution. Our tasks require models to infer task-relevant physical structure and generate feasible action sequences under explicit constraints, including contact-rich manipulation, tool use, and multi-stage dependencies. We introduce a benchmark of 35 tasks, 14K filtered trajectories, and scalable tools for generating diverse scenarios. Experiments reveal a consistent gap: vision language models show partial physical reasoning ability but fail to produce executable plans, while state-of-the-art vision-language-action models struggle to satisfy task constraints and generalize across scenarios. These results identify intuitive manipulation as a missing axis in current foundation models and generalist robot policies, and position IMBENCH as a step toward evaluating and enabling more integrated, adaptive physical intelligence.

Problem

• intuitive manipulation
• physical reasoning
• robotic manipulation
• benchmark
• foundation models

Innovation

• intuitive manipulation
• physical reasoning
• action generation
• robotic benchmark
• vision-language-action models