Qwen2.5-Omni-7B: benchmark results
Alibaba's 7B end-to-end omni-modal model (March 2025) whose Thinker-Talker design takes text, image, audio, and video and streams speech replies. Provider: Alibaba. Released 2025-03-26. Access: Open.
Aggregate Rating 1534 ± 7, rank #1797 of 10870 rated models, from 672 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Afrispeech Semantics - Consistency (AfriSpeech-General) | 91.31 | Macro-F1 (%; binary consistent versus inconsistent judgments | 100 |
| Afrispeech Semantics - Plausibility (AfriSpeech-General) | 95.69 | Macro-F1 (%; plausible-but-unsupported versus implausible ju | 100 |
| Audio2Tool - Conversation | 26.1 | Exact match (%): share of queries whose predicted tool or or | 100 |
| Audio2Tool - Correction | 42 | Exact match (%): share of queries whose predicted tool or or | 100 |
| Audio2Tool - Multi-Intent | 32.1 | Exact match (%): share of queries whose predicted tool or or | 100 |
| Audio2Tool - Needle-in-a-Haystack | 40.2 | Exact match (%): share of queries whose predicted tool or or | 100 |
| Chehre - Distributional Coverage (High Ambiguity, Random Sampling) | 76 | Coverage (%; share of the human annotators' rating vectors t | 100 |
| Chehre - Distributional Coverage (Low Ambiguity, Random Sampling) | 65 | Coverage (%; share of the human annotators' rating vectors t | 100 |
| HELM Audio - Air Bench Foundation | 74.35 | EM | 100 |
| HELM Audio - Vocal Sound | 90.35 | PEM | 100 |
| HUG-VIS - Emotion Recognition (Video + Audio + Text) | 83.79 | Seven-class emotion recognition accuracy (%; 8,400 seated ha | 100 |
| HUG-VIS - Emotion Recognition (Video + Audio) | 73.7 | Seven-class emotion recognition accuracy (%; 8,400 seated ha | 100 |
Interactive version: theaggregate.ai/model?slug=qwen2-5-omni-7b · How It Works · Data refreshed daily, snapshot 2026-11-12.