MiDashengLM-7B: benchmark results
Provider: Other.
Aggregate Rating 1573 ± 45, rank #1363 of 10916 rated models, from 15 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| RAIL - Auditory Processing | 46.92 | LLM-as-judge score (%) on auditory-processing tasks (speech- | 72 |
| MCGA - ASR - Wen | 10 | Character error rate (%; 200 recorded wen prose passages of | 60 |
| RAIL - Memory | 64.08 | LLM-as-judge score (%) on auditory memory tasks (memory span | 56 |
| MCGA | 326.2 | Score (0-600; sum of the six task scores on 1,000 test recor | 55.6 |
| MCGA - Speech Reasoning | 75.6 | Accuracy (%; multiple-choice questions on 1,000 test recordi | 55.6 |
| MCGA - Speech Translation | 42.9 | LLM-B (0-100; mean of beauty of form, meaning and sound of t | 55.6 |
| RAIL - Fluid Reasoning | 51.24 | LLM-as-judge score (%) on fluid-reasoning tasks (induction, | 54 |
| MCGA - ASR - Ci | 10.1 | Character error rate (%; 200 recorded ci lyrics of the test | 50 |
| MCGA - ASR - Fu | 15.7 | Character error rate (%; 200 recorded fu rhapsodies of the t | 50 |
| MCGA - ASR - Qu | 9.4 | Character error rate (%; 200 recorded qu verses of the test | 50 |
| MCGA - ASR - Shi | 12.7 | Character error rate (%; 200 recorded shi poems of the test | 50 |
| MCGA - Speech Emotion Captioning | 24.7 | Emotion Caption Fidelity (0-100; persona, global tone and se | 44.4 |
Interactive version: theaggregate.ai/model?slug=midashenglm-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.