MiMo-VL-7B-SFT: benchmark results
Provider: Xiaomi. Access: Open.
Aggregate Rating 1590 ± 18, rank #1123 of 10916 rated models, from 61 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| TSHA | 80.4 | Final average (%): mean of the open-ended QA overall (GPT-4o | 90.5 |
| VISTA-Bench - Multimodal Reasoning | 61.3 | Accuracy (%) on the 300 multimodal reasoning questions (logi | 89.7 |
| VISTA-Bench | 62.2 | Accuracy (%) over all 1,500 questions (300 perception, 300 r | 86.2 |
| VISTA-Bench - Multimodal Knowledge | 46.5 | Accuracy (%) on the 400 multimodal knowledge questions (STEM | 86.2 |
| VISTA-Bench - Unimodal Knowledge | 71.4 | Accuracy (%) on the 500 unimodal knowledge questions (natura | 86.2 |
| Artifact-Bench (AI-Generated Video) - Pairwise Realism Comparison | 49.8 | Accuracy (%) on the 250 questions asking which of two AI-gen | 83.3 |
| E-VAds (Relaxed-5) - Cross-Modal Detection | 67.6 | Relaxed-5 score (%) on the Cross-Modal Detection task (consi | 81.2 |
| E-VAds (Relaxed-5) - Regulatory Compliance | 61.6 | Relaxed-5 score (%) on the Regulatory Compliance task (poten | 81.2 |
| TSHA - Choice Questions | 85.5 | Choice-question overall (%): mean of yes/no hazard-presence | 81 |
| TSHA - Open-Ended Hazard QA | 75.2 | Open-ended QA overall (%): GPT-4o rubric score, 0.7 hazard a | 81 |
| VISTA-Bench - Multimodal Perception | 68.7 | Accuracy (%) on the 300 multimodal perception questions (glo | 77.6 |
| SIS-Bench - Spatial Consistency | 71.3 | Accuracy (%; 195 spatial consistency questions; four-option | 76 |
Interactive version: theaggregate.ai/model?slug=mimo-vl-7b-sft · How It Works · Data refreshed daily, snapshot 2026-11-08.