TimeLogic Challenge @ CVPR 2026: Strong MLLMs Meet Evidence-Seeking Agents for Temporal-Logic Video Question Answering
For researchers in video understanding and temporal reasoning, this work addresses the limitation of single-pass models on temporally localized evidence, offering a strong benchmark result.
The authors tackled temporal-logic video question answering, where models must reason about action timing (before, after, etc.), and proposed a training-free evidence-seeking agent that uses a Think-Act-Observe loop with multi-granular sampling and timestamp interleaving, achieving 77.13% average accuracy on the TimeLogic test set.
Temporal-logic video question answering requires a model to reason about when actions occur relative to one another, such as before, after, until, since, overlap, and multi-event chains, rather than merely what is present in a video. Standard vision-language models typically answer such questions in a single pass over a fixed, uniformly sampled set of frames, which is poorly matched to evidence that is often localized to narrow action boundaries or dispersed across several distant events. We present an evidence-seeking agent that treats temporal-logic VideoQA as active exploration. The agent follows a Think-Act-Observe loop driven by a multi-granular sampling toolkit, where every observation is interleaved with its absolute timestamp so that temporal relations reduce to numerical comparisons on a shared time axis. Its behavior is shaped by benchmark structure: a lightweight classifier routes each question to a temporal category, each with a tailored policy, iteration depth, and prompt, while sampling budgets adapt to corpus characteristics and clip length. The resulting training-free system couples Gemini 3.1 Pro with a temporal-reasoning policy and achieves 77.13 AvgAcc on the official TimeLogic test set.