Codesota · Computer Vision · Few-Shot Image Classification · SciVideoBenchTasks/Computer Vision/Few-Shot Image Classification
Few-Shot Image Classification · benchmark dataset · ENGLISH

SciVideoBench: Benchmarking Scientific Video Reasoning in Large Multimodal Models.

The first comprehensive benchmark dedicated to scientific video reasoning. SciVideoBench evaluates models across Physics, Chemistry, Biology, and Medicine, covering both perceptual understanding and high-level reasoning tasks. It provides a rigorous benchmark for evaluating long-form video reasoning in domains where accuracy and explainability matter most. Features 1,000 high-quality, human-verified multiple-choice questions across 240+ scientific experiments with rich metadata including discipline, subject, timestamp breakdowns, and rationale.

Paper Submit a result
§ 01 · Leaderboard

Best published scores.

No results indexed yet — be the first to submit a score.

No benchmark results indexed yet
§ 06 · Contribute

Have a score that beats
this table?

Submit a checkpoint and a reproduction script. We will run it, publish the score, and — if it takes the top — annotate the step on the progress chart with your name.

Submit a result Read submission guide
What a submission needs
  • 01A public checkpoint or API endpoint
  • 02A reproduction script with frozen commit + seed
  • 03Declared evaluation environment (Python, deps)
  • 04One row per metric declared by this dataset
  • 05A contact so we can follow up on discrepancies
SciVideoBench — Few-Shot Image Classification benchmark · Codesota | CodeSOTA