Ming-Flash-Omni-2.0: benchmark results
Provider: Other.
Aggregate Rating 1578 ± 33, rank #1280 of 10916 rated models, from 39 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| TraceAV-Bench (Audio Only) | 30.78 | Accuracy (%), macro-average of the 12 general sub-tasks, fou | 100 |
| TraceAV-Bench - Visual Counting | 39.38 | Accuracy (%) on the 226 Visual Counting (VC) questions, four | 82.1 |
| TraceAV-Bench - Cross-Modality Matching | 44.71 | Accuracy (%) on the 85 Cross-Modality Matching (CMM) questio | 78.6 |
| TraceAV-Bench | 51.7 | Accuracy (%), macro-average of the 12 general sub-tasks, fou | 71.4 |
| TraceAV-Bench - Background Music | 63.36 | Accuracy (%) on the 131 Background Music (BM) questions, fou | 71.4 |
| TraceAV-Bench - Spatial Reasoning | 65.45 | Accuracy (%) on the 165 Spatial Reasoning (SR) questions, fo | 71.4 |
| TraceAV-Bench - Temporal Sequencing | 53.61 | Accuracy (%) on the 97 Temporal Sequencing (TS) questions, f | 71.4 |
| TraceAV-Bench (Visual Only) | 38.55 | Accuracy (%), macro-average of the 12 general sub-tasks, fou | 66.7 |
| TraceAV-Bench - Entity Tracking | 47.58 | Accuracy (%) on the 124 Entity Tracking (ET) questions, four | 64.3 |
| TraceAV-Bench - Information Retrieval | 56.43 | Accuracy (%) on the 140 Information Retrieval (IR) questions | 64.3 |
| TraceAV-Bench - Multimodal Hallucination | 67.07 | Accuracy (%), average of the three multimodal hallucination | 64.3 |
| TraceAV-Bench - Visual-to-Audio Deception | 71.3 | Accuracy (%) on the 230 Visual-to-Audio Deception (V2A) ques | 64.3 |
Interactive version: theaggregate.ai/model?slug=ming-flash-omni-2-0 · How It Works · Data refreshed daily, snapshot 2026-11-08.