ERNIE 4.5 VL 28B A3B: benchmark results
Provider: Baidu. Access: API.
Unified ELO 1572 ± 35, rank #742 of 2656 rated models, from 13 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SWE-Arena | 1002 | Elo Score | 80.6 |
| GEB-Bench - Story-Theorem Matching | 41.3 | Accuracy (%; 4-way, chance 25%) | 52.8 |
| GEB-Bench - Scene-Story Matching | 32.4 | Accuracy (%; 4-way, chance 25%) | 46.2 |
| GEB-Bench - Skeleton Identification | 46.9 | Accuracy (%; 25-way, chance 4%) | 30.8 |
| GEB-Bench - Scene-Theorem Matching | 42.9 | Accuracy (%; 4-way, chance 25%) | 26.9 |
| VisuLogic | 26 | Overall Accuracy (%) | 26 |
| CFMME | 45.68 | Average (self-reported) | 20 |
| GEB-Bench - Adversarial Matching | 57.1 | Accuracy (%; 4-way, chance 25%) | 19.2 |
| GEB-Bench - Scene Identification | 30.5 | Accuracy (%; 25-way, chance 4%) | 15.4 |
| SnakeBench | 13.9 | TrueSkill Rating | 14.8 |
| GEB-Bench - Average (Vision Models) | 35.6 | Macro-average accuracy over all eight tasks (%) | 11.5 |
| GEB-Bench - Story Identification | 14 | Accuracy (%; 25-way, chance 4%) | 6.9 |
Interactive version: theaggregate.ai/model?slug=ernie-4-5-vl-28b-a3b · How It Works · Data refreshed daily, snapshot 2026-09-19.