Visual Question Answering on BenchLMM
Metric: GPT-3.5 score (higher is better)
Metric: GPT-3.5 score (higher is better)
| # | Model↕ | GPT-3.5 score▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | GPT-4V | 58.37 | Yes | GPT-4 Technical Report | 2023-03-15 | Code |
| 2 | Sphinx-V2-1K | 57.43 | Yes | SPHINX: The Joint Mixing of Weights, Tasks, and ... | 2023-11-13 | Code |
| 3 | LLaVA-1.5-13B | 55.53 | No | Improved Baselines with Visual Instruction Tuning | 2023-10-05 | Code |
| 4 | LLaVA-1.5-7B | 46.83 | No | Visual Instruction Tuning | 2023-04-17 | Code |
| 5 | InstructBLIP-13B | 45.03 | No | InstructBLIP: Towards General-purpose Vision-Lan... | 2023-05-11 | Code |
| 6 | InstructBLIP-7B | 44.63 | No | InstructBLIP: Towards General-purpose Vision-Lan... | 2023-05-11 | Code |
| 7 | LLaVA-1-13B | 43.5 | No | Visual Instruction Tuning | 2023-04-17 | Code |
| 8 | Otter-7B | 39.13 | No | Otter: A Multi-Modal Model with In-Context Instr... | 2023-05-05 | Code |
| 9 | MiniGPT4-13B | 34.93 | No | MiniGPT-4: Enhancing Vision-Language Understandi... | 2023-04-20 | Code |
| 10 | MiniGPTv2-7B | 30.1 | No | MiniGPT-v2: large language model as a unified in... | 2023-10-14 | Code |