Visual Question Answering on RGB-Th Bench
79.68RGB-Txt AccuracyInternVL 2.5-8B (IVL)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| InternVL 2.5-8B (IVL)LLM backbone=InternLM2.5-7B, Vision backbone=InternViT-300M-v2.52026.05 | 79.68 | 63.31 | 71.5 | |
| LLaVA-Next-Mistral-7.6B (LlaVANext)LLM backbone=Mistral-7B, Vision backbone=CLIP ViT-L/142026.05 | 77.83 | 52.82 | 65.33 | |
| Molmo-7B-D-0924 (Molmo)LLM backbone=Qwen2-7B, Vision backbone=CLIP ViT-L/142026.05 | 77.25 | 60.84 | 69.05 | |
| Thermo-VLLLM backbone=Qwen2-7B, Vision backbone=CLIP-ViT-L/142026.05 | 77.25 | 69.15 | 73.2 | |
| LLaVA-Next-Llama3-8B (LlaVANext)LLM backbone=Llama-3 8B, Vision backbone=CLIP ViT-L/142026.05 | 77.22 | 58.63 | 67.92 | |
| Qwen2.5-VL-7B-Instruct (QwenVL)LLM backbone=Qwen2.5-7B, Vision backbone=QwenViT2026.05 | 75.62 | 67.26 | 71.44 |