AI BENCHMARK PROFILE

Multi2AV-Safety

General AISafety & Trustworthiness

Audio-video generation is rapidly moving from prompt-driven synthesis toward multimodal conditioning, where text, images, audio, and video can jointly shape the generated output.

Released: 2026-08-27
Readiness: Paper only
Primary field: General AI

Motivation

Audio-video generation is rapidly moving from prompt-driven synthesis toward multimodal conditioning, where text, images, audio, and video can jointly shape the generated output.

Primary resources

Paper ↗Hugging Face ↗

Benchmark Radar records only publicly supported details and links back to primary sources for verification.