Seed1.6-VL: benchmark results
Provider: ByteDance.
Aggregate Rating 1694 ± 28, rank #377 of 10916 rated models, from 32 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| VURB - Long Video Understanding | 69 | Majority-vote pairwise accuracy (%; eight judgments with the | 100 |
| M-JudgeBench - Incidental Mistake Detection | 98.02 | pairwise accuracy (%) on the 252 pairs where the rejected ch | 94.7 |
| M-JudgeBench - Logical Fallacy Detection | 98.74 | pairwise accuracy (%) on the 238 pairs where the rejected ch | 94.7 |
| VURB | 62.6 | Majority-vote pairwise accuracy (%; eight judgments with the | 92.9 |
| VURB - General Video Understanding | 61.6 | Majority-vote pairwise accuracy (%; eight judgments with the | 92.9 |
| M-JudgeBench - Long CoT Same-Model Pairs | 60.53 | pairwise accuracy (%) on the 337 pairs where a correct and a | 92.1 |
| M-JudgeBench - Visual Perception Error Detection | 97.58 | pairwise accuracy (%) on the 248 pairs where the rejected ch | 92.1 |
| ChartJudgeBench - CPA Color Consistency | 91.5 | Pairwise judging accuracy (%; the 117 color consistency pair | 92 |
| ChartJudgeBench - CPA Text Occlusion | 69.4 | Pairwise judging accuracy (%; the 157 text occlusion pairs; | 92 |
| ChartJudgeBench - CPA Visual Style | 68.2 | Pairwise judging accuracy (%; the 88 visual style pairs; two | 92 |
| M-JudgeBench | 55.33 | Overall pairwise accuracy (%), pooled over all ten subtasks, | 89.5 |
| ChartJudgeBench - Chart Perception Alignment | 64.5 | Pairwise judging accuracy (%; unweighted mean of the seven f | 88 |
Interactive version: theaggregate.ai/model?slug=seed1-6-vl · How It Works · Data refreshed daily, snapshot 2026-11-08.