Benchmark
MAEB(v1, audio-only) - Mean (TaskType)
Scores are percentages.
- Models measured
- 8
- Scores recorded
- 8
- Sources
- 1
- IQ weight
- 1.23
Score distribution
- 51%
- Mean
- 48.9%
- Median
- 4.1%
- Std dev
- 48.2%
- 10th percentile
- 57.7%
- 90th percentile
- 47.7%
- Min
This benchmark has a weight of 1.23 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 | Jina Embeddings v5 Omni SmallMean (TaskType)· MTEB Multimodal Boards data docs | 58.4% |
| 2 | Jina Embeddings v5 Omni NanoMean (TaskType)· MTEB Multimodal Boards data docs | 57.7% |
| 3 | Whisper MediumMean (TaskType)· MTEB Multimodal Boards data docs | 50.7% |
| 4 | embeddinggemma-2Mean (TaskType)· MTEB Multimodal Boards data docs | 48.9% |
| 5 | Whisper BaseMean (TaskType)· MTEB Multimodal Boards data docs | 48.3% |
| 6 | Whisper Large V3Mean (TaskType)· MTEB Multimodal Boards data docs | 48.2% |
| 7 | Whisper SmallMean (TaskType)· MTEB Multimodal Boards data docs | 48.2% |
| 8 | Whisper TinyMean (TaskType)· MTEB Multimodal Boards data docs | 47.7% |