Step-1.5V: benchmark results
Provider: StepFun. Released 2024-07-04. Access: API.
Aggregate Rating 1492 ± 46, rank #2481 of 10916 rated models, from 259 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM CCBench - Scenery and Buildings | 95.8 | Accuracy (%) | 99.1 |
| OpenVLM LLaVA-Bench - Detail | 96.6 | Score | 98.5 |
| OpenVLM MMBench V1.1 CN - Celebrity Recognition | 95.2 | Accuracy (%) | 98.4 |
| OpenVLM MMBench V1.1 EN - Celebrity Recognition | 95.9 | Accuracy (%) | 98.4 |
| OpenVLM AI2D - Food Chains and Webs | 98.1 | Accuracy (%) | 98.2 |
| OpenVLM MMMU - Diagnostics and Laboratory Medicine | 56.7 | Accuracy (%) | 98.2 |
| OpenVLM SEEDBench IMG - Instance Location | 73.9 | Accuracy (%) | 97.8 |
| OpenVLM MMMU - Clinical Medicine | 76.7 | Accuracy (%) | 97.7 |
| OpenVLM MMMU - Art and Design | 75 | Accuracy (%) | 97.5 |
| OpenVLM OCRBench - Scene Text-centric VQA | 95.5 | Correct Answers (%) | 97.5 |
| OpenVLM MMMU - Music | 46.7 | Accuracy (%) | 96.8 |
| OpenVLM MMMU - Energy and Power | 66.7 | Accuracy (%) | 96.6 |
Interactive version: theaggregate.ai/model?slug=step-1-5v · How It Works · Data refreshed daily, snapshot 2026-11-08.