Flash-VStream-7B: benchmark results
Provider: Other.
Aggregate Rating 1360 ± 51, rank #7449 of 10916 rated models, from 33 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SVCBench - O1-Delta | 45.9 | Gaussian Precision Accuracy (0-100): per query point exp(-(p | 92.3 |
| SVCBench - O2-Gain | 56.4 | Gaussian Precision Accuracy (0-100): per query point exp(-(p | 92.3 |
| SVCBench - E1-Action | 19 | Gaussian Precision Accuracy (0-100): per query point exp(-(p | 42.3 |
| SVCBench | 23.4 | GPA (Global Performance Average, %) | 38.5 |
| OVBench | 31.2 | AVG | 26.7 |
| SVCBench - E2-Episode | 17.2 | Gaussian Precision Accuracy (0-100): per query point exp(-(p | 23.1 |
| SVCBench - O2-Unique | 21.9 | Gaussian Precision Accuracy (0-100): per query point exp(-(p | 23.1 |
| IPIBench - Task Cancellation | 2.31 | Accuracy (%) on task cancellation (stop responding after the | 20 |
| RealStreamEval - Forward Active Responding | 12 | Score (%; RealStreamEval on OVO-Bench tasks and questions: f | 20 |
| SVCBench - E1-Transit | 11.9 | Gaussian Precision Accuracy (0-100): per query point exp(-(p | 15.4 |
| IPIBench - Reactive-under-Proactive | 0 | Accuracy (%) on reactive-under-proactive interaction (answer | 10 |
| RIVER Bench - Live-Perception | 29.28 | Multiple-choice accuracy (%) on RIVER Bench Live-Perception | 8.3 |
Interactive version: theaggregate.ai/model?slug=flash-vstream-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.