MotionHalluc: Diagnosing Kinematic Hallucinations in Fine-Grained Motion Reasoning

Video Large Language Models (VideoLLMs) have shown strong progress in video understanding, yet they still suffer from hallucinations that are inconsistent with visual evidence. Existing benchmarks mainly focus on object hallucination or coarse action perception, leaving a key video-specific problem underexplored: motion hallucination, in which models infer human motions that are absent from the video. We present MoHallBench, a benchmark for diagnosing motion hallucination in VideoLLMs. MoHallBench systematically evaluates three major sources of hallucination: co-occurrence priors, sequential inference, and similarity confusion. It contains 11,306 video clips and 40,493 question-answer pairs, covering binary-choice, multiple-choice, and generative settings. We further introduce a bi-directional questioning protocol with bias-aware metrics to reduce affirmation bias in binary evaluation. Experiments on ten recent open-source VideoLLMs reveal a clear decoupling between action recognition and hallucination resistance, as models that perform well on positive action recognition often fail on adversarial negatives. Among all settings, sequential inference hallucination is the most severe, showing that current models tend to over-infer expected outcomes from partial motion cues. Our analyses further confirm that stronger priors and finer-grained similarity substantially amplify hallucination. We hope MoHallBench can facilitate future evaluation and mitigation of motion hallucination in VideoLLMs.

We introduce VidPair-Halluc, a new benchmark for evaluating video hallucination in large video models (LVMs) under rigorous and controlled conditions. Unlike previous benchmarks that primarily rely on text-based perturbations or adversarial questions while neglecting the consistency of visual backgrounds, VidPair-Halluc features video pairs with highly similar backgrounds but distinctly different foreground semantics, enabling precise attribution of model errors to genuine hallucination rather than background variation. The benchmark is constructed through PairFlow, a pipeline that leverages recent advances in text-to-image and video generation to systematically compose stories, generate coherent video clips, and assemble them into adversarial pairs. Covering both spatial and temporal reasoning across ten semantic aspects, VidPair-Halluc comprises 1K high-quality adversarial video pairs and 11K spatio-temporal QA pairs with control over background and foreground variations. Evaluations on mainstream LVMs show persistent difficulty with robust fine-grained video understanding in adversarial settings, and code and data are available at the https://jethrojames.github.io/VidPair-Halluc/.

Video large language models (Video-LLMs) are increasingly used as the perceptual front end of world models, a role that assumes they can read motion: how fast something moves, which way it travels, how hard it is pushed. We show they cannot. A Video-LLM can watch two clips of the same person in the same room, name every object in both, and still fail to say which clip moves faster. We introduce MotionBlind, a contrastive benchmark of self-recorded video for physically grounded motion(speed, magnitude, and direction), the variables a world model must predict. Each instance is a pair of near-identical clips that differ only in motion. Each clip carries two complementary yes/no questions, giving four items per instance, and a model earns credit only if all four are correct. We report Instance Accuracy(IAcc), which has a 6.25% chance floor. Single-frame, appearance, and language-only shortcuts all collapse to it. MotionBlind complements the recent TimeBlind benchmark. We run a controlled study of six open and two frontier Video-LLMs, varying whether the video is present, whether frames are shown in the correct temporal order, and how frames are sampled (1 to 24 frames, four selection strategies). Open models sit near the 6.25% floor, and scale does not help. Removing the video drops every model to zero IAcc, and shuffling frames collapses IAcc to chance, so the task genuinely needs video in order. Neither more frames nor smarter frame selection closes the gap, because these change which frames are seen, not whether motion is read. Only Gemini3.1 Pro clears the benchmark overall, and even it fails on speed. A frontend that cannot tell two speeds of the same action apart is not yet a trustworthy source of supervision, reward, or evaluation for a world model.