qwen3-omni-30B: benchmark results

Provider: Alibaba. Released 2025-09-22. Access: Open.

Aggregate Rating 1549 ± 34, rank #1608 of 10516 rated models, from 37 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
TempCompass72.35Avg All (%)92.9
AudioRAG37Accuracy (%) on AudioRAG's 500 multi-hop questions about a s80
ESDR-Bench63.83Pairwise preference accuracy (%) over all benchmark pairs (m77.8
ESDR-Bench - Modality58.18Pairwise preference accuracy (%) on the modality pairs (micr77.8
ESDR-Bench - Colloquialness97.2Pairwise preference accuracy (%) on the colloquialness pairs66.7
MSI-Bench - English Selective Disclosure57.3All-pass rate (%; the 96 selective-disclosure cases, where e64.3
MSI-Bench - English34All-pass rate (%; all 576 English test cases over six multi-57.1
MSI-Bench - English Sequential Constraints27.1All-pass rate (%; the 96 sequential-constraint cases, where 57.1
MSI-Bench - English Speaker Authority63.5All-pass rate (%; the 96 speaker-authority cases, where an u57.1
MSI-Bench - Mandarin19.3All-pass rate (%; all 576 Mandarin test cases over six multi57.1
MSI-Bench - Mandarin Constraint Prioritization35.4All-pass rate (%; the 96 constraint-prioritization cases, wh57.1
MSI-Bench - Mandarin Sequential Constraints17.7All-pass rate (%; the 96 sequential-constraint cases, where 57.1

Interactive version: theaggregate.ai/model?slug=qwen3-omni-30b · How It Works · Data refreshed daily, snapshot 2026-11-21.