Benchmark
MAEB(v1) - AudioPairClassification
Scores are percentages.
- Models measured
- 11
- Scores recorded
- 11
- Sources
- 1
- IQ weight
- 0.83
Score distribution
- 56.2%
- Mean
- 52.9%
- Median
- 6.1%
- Std dev
- 51.5%
- 10th percentile
- 65.6%
- 90th percentile
- 50.8%
- Min
This benchmark has a weight of 0.83 for speech models, measured over 5 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | embeddinggemma-2AudioPairClassification· MTEB Multimodal Boards data docs | 67.5% |
| 2 | Jina Embeddings v5 Omni NanoAudioPairClassification· MTEB Multimodal Boards data docs | 65.6% |
| 3 | Jina Embeddings v5 Omni SmallAudioPairClassification· MTEB Multimodal Boards data docs | 65.4% |
| 4 | Whisper MediumAudioPairClassification· MTEB Multimodal Boards data docs | 53.9% |
| 5 | qwen3-omni-30b-a3b-instructAudioPairClassification· MTEB Multimodal Boards data docs | 53.9% |
| 6 | Qwen2.5-Omni 7BAudioPairClassification· MTEB Multimodal Boards data docs | 52.9% |
| 7 | Whisper SmallAudioPairClassification· MTEB Multimodal Boards data docs | 52.6% |
| 8 | Whisper Large V3AudioPairClassification· MTEB Multimodal Boards data docs | 52.5% |
| 9 | Whisper BaseAudioPairClassification· MTEB Multimodal Boards data docs | 52.1% |
| 10 | Whisper TinyAudioPairClassification· MTEB Multimodal Boards data docs | 51.5% |
| 11 | Qwen3 Omni 30B A3B ThinkingAudioPairClassification· MTEB Multimodal Boards data docs | 50.8% |