BAGEL-7B-MoT: benchmark results
Provider: Jon Durbin. Access: Open.
Aggregate Rating 1411 ± 36, rank #5579 of 10916 rated models, from 176 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| DLEBench | 61 | Average score (% of the maximum): mean of instruction follow | 88.9 |
| XTC-Bench - Attribute Query | 70.4 | Understanding score (0-1, times 100) on attribute questions | 88.9 |
| XTC-Bench - Relation Query | 75.9 | Understanding score (0-1, times 100) on relation questions b | 88.9 |
| XTC-Bench - Understanding | 72.4 | Overall understanding score (0-1, times 100) on visual quest | 88.9 |
| UniEditBench (Image) - Background Consistency | 4.38 | Background consistency score (1-5): regions outside the edit | 87.5 |
| XTC-Bench | 56.9 | Accuracy-weighted continuous cross-task agreement (AW-CCTA, | 87.5 |
| Unison - Internal Consistency | 80.3 | Unified score (%) on the internal consistency dimension of U | 85.7 |
| XTC-Bench - Generation | 72.5 | Overall generation score (0-1, times 100): the model draws a | 77.8 |
| AIM-Bench (Affective Image Manipulation) - Negative Emotion Accuracy | 56.61 | Emotion accuracy (%) on the items whose target is a negative | 75 |
| TinyGrafixBench - Line Chart | 4.9 | mIoU (%) | 75 |
| MedQ-Deg - Imaging Perception | 32.5 | Accuracy (%) on the Imaging Perception questions of MedQ-Deg | 74.4 |
| SpatialGen-Bench (Text Answering) | 53.96 | Overall Accuracy (%) | 73.1 |
Interactive version: theaggregate.ai/model?slug=bagel-7b-mot · How It Works · Data refreshed daily, snapshot 2026-11-08.