EXAONE 4.5 33B: benchmark results
LG AI Research's open 33B vision-language model with hybrid attention (research license). Provider: LG AI. Released 2026-04-09. Access: Open.
Unified ELO 1571 ± 1, rank #361 of 1392 rated models, from 42 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| K-MetBench | 75.9 | Accuracy (self-reported) | 81 |
| CritPt | 30 | Accuracy (self-reported) | 80.8 |
| AA IFBench | 57.96 | Accuracy (%) | 69.8 |
| AA TAU-2 Bench | 78.07 | Accuracy (%) | 68.5 |
| LLM Stats (MathVista-Mini) | 85 | Score (%) | 66.7 |
| AA GPQA Diamond | 79.39 | Accuracy (%) | 66.3 |
| AA Humanity's Last Exam | 12.88 | Accuracy (%) | 62.2 |
| ZeroEval GPQA Diamond | 80.5 | GPQA Diamond Score | 61.5 |
| LLM Stats Score | 26.28 | LLM Stats Score (conservative rating) | 60.1 |
| AA Terminal-Bench Hard | 20.45 | Accuracy (%) | 59.9 |
| LLM Stats (AI2D) | 89 | Score (%) | 57.6 |
| BenchmarkList ECI | 116.12 | Capability Index (ECI) | 56.3 |
Interactive version: theaggregate.ai/model?slug=exaone-4-5-33b · How It Works · Data refreshed daily, snapshot 2026-09-05.