| Benchmark | Model | Rank | Results |
|---|---|---|---|
| natural-language-visual-grounding-on | CogAgent | #13 | Accuracy (%): 47.4 |
| visual-question-answering-on-mm-vet | CogAgent | #40 | GPT-4 score: 52.8Params: 18B |
| visual-question-answering-on-mm-vet-v2 | CogAgent-Chat | #13 | GPT-4 score: 34.7±0.2 |