LLaVA-NeXT-13B: benchmark results
Provider: LLaVA Team.
Aggregate Rating 1403 ± 46, rank #6056 of 10489 rated models, from 27 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CAGE - Counterfactual (Chain-Text Probe) - Chain Score | 7.61 | Judge score (0-10) for the causal chain in arrow notation, m | 100 |
| CAGE - Counterfactual (Chain-Text Probe) - Text Score | 7.64 | Judge score (0-10) for the textual answer, mean of GPT-4o an | 100 |
| CAGE - Intervention (Chain-Text Probe) - Chain Score | 7.83 | Judge score (0-10) for the causal chain in arrow notation, m | 100 |
| CAGE - Intervention (Chain-Text Probe) - Text Score | 7.96 | Judge score (0-10) for the textual answer, mean of GPT-4o an | 100 |
| EIBench (complex) | 39.16 | Recall | 91.7 |
| EIBench | 54.33 | Recall | 83.3 |
| PlantMicro - Host | 40.1 | Host plant species identification accuracy (%); four-option | 80 |
| PlantMicro - Localization mAP@0.5 | 35.5 | Mean average precision at IoU 0.5 (%) of directly predicted | 75 |
| CAGE - Counterfactual (Text-Only Probe) | 7.94 | Judge score (0-10), mean of GPT-4o and Claude 3.5 Sonnet rat | 71.4 |
| CAGE - Intervention (Text-Only Probe) | 8.08 | Judge score (0-10), mean of GPT-4o and Claude 3.5 Sonnet rat | 71.4 |
| CAGE - Association (Text-Only Probe) | 7.95 | Judge score (0-10), mean of GPT-4o and Claude 3.5 Sonnet rat | 57.1 |
| PlantMicro - Pathogen | 29.56 | Pathogen identification accuracy (%); four-option multiple c | 50 |
Interactive version: theaggregate.ai/model?slug=llava-next-13b · How It Works · Data refreshed daily, snapshot 2026-11-20.