Codesota · General · Omni models · DailyOmniTasks/General/Omni models
Omni models · benchmark dataset · EN

Daily-Omni: Towards Audio-Visual Reasoning with Temporal Alignment across Modalities.

Daily-Omni is an audio–visual question-answering benchmark for audio-visual reasoning that emphasizes temporal alignment across modalities. According to the dataset repository and paper listing, Daily-Omni comprises real-world short videos of daily-life scenarios and multiple-choice QA pairs designed to require integration of audio and visual streams. The project provides a QA generation pipeline (automatic annotation, QA generation and QA optimization) to scale creation and human evaluation, and includes a baseline agent (Daily-Omni-Agent) combining open-source visual-language, audio-language and ASR models with simple temporal alignment methods. The dataset listing on Hugging Face and the project repository report the benchmark contains 684 videos and 1,197 multiple-choice QA pairs across six main task categories (question-answering tasks focused on audio-visual integration).

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