Skywork-R1V3-38B: benchmark results
Provider: Skywork. Released 2025-07-09. Access: Open.
Aggregate Rating 1613 ± 21, rank #898 of 10916 rated models, from 37 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMMU Benchmark | 76 | Validation Score | 88.8 |
| HAERAE-Vision - Machinery | 37.25 | Checklist Score (%; original under-specified queries, GPT-5- | 77.3 |
| HAERAE-Vision - Health/Medical | 47.94 | Checklist Score (%; original under-specified queries, GPT-5- | 75 |
| HAERAE-Vision - Natural Objects | 27.38 | Checklist Score (%; original under-specified queries, GPT-5- | 75 |
| MME-Reasoning | 33 | Overall Accuracy (%) | 75 |
| HAERAE-Vision - IT/Computer | 27.12 | Checklist Score (%; original under-specified queries, GPT-5- | 72.7 |
| HAERAE-Vision - Transportation | 30.1 | Checklist Score (%; original under-specified queries, GPT-5- | 72.7 |
| HAERAE-Vision | 27.76 | Checklist Score (%; original under-specified queries, GPT-5- | 70.5 |
| HAERAE-Vision - Business/Economics | 32.37 | Checklist Score (%; original under-specified queries, GPT-5- | 70.5 |
| HAERAE-Vision - Entertainment/Arts | 14.76 | Checklist Score (%; original under-specified queries, GPT-5- | 70.5 |
| HAERAE-Vision - Gaming | 15.3 | Checklist Score (%; original under-specified queries, GPT-5- | 70.5 |
| HAERAE-Vision - Mathematics | 41.71 | Checklist Score (%; original under-specified queries, GPT-5- | 70.5 |
Interactive version: theaggregate.ai/model?slug=skywork-r1v3-38b · How It Works · Data refreshed daily, snapshot 2026-11-08.