Science Question Answering on ScienceQA IMG (test)
98.4AccuracyInternVL2.5-8B
Evaluation results
| Method | Links | |
|---|---|---|
| InternVL2.5-8B2026.04 | 98.4 | |
| InternVL2-8B2026.04 | 97.1 | |
| MiniCPM-V-2.6-8B2026.04 | 96.7 | |
| IXC-2.52026.04 | 96.6 | |
| LLaVA-OneVision-7B2026.04 | 95.4 | |
| Qwen2.5-VL-7B + Saliency-R1Training Stage=Saliency-R12026.04 | 94.3 | |
| Qwen2.5-VL-7B + SFTTraining Stage=SFT2026.04 | 93.8 | |
| Qwen2.5-VL-3B + Saliency-R1Training Stage=Saliency-R12026.04 | 91.2 | |
| Qwen2.5-VL-3B + SFTTraining Stage=SFT2026.04 | 90.5 | |
| GPT-4o2026.04 | 90.1 | |
| MiniCPM-Llama-V-2.5-8B2026.04 | 89.2 | |
| Claude-3.5 Sonnet2026.04 | 88.9 | |
| Qwen2.5-VL-7BTraining Stage=Base2026.04 | 88.2 | |
| Gemini-1.5-Pro2026.04 | 85.7 | |
| GPT-4V-1106Size=> 8B2024.02 | 82.1 | |
| Qwen2.5-VL-3BTraining Stage=Base2026.04 | 81.2 | |
| Cambrain-1-8B2026.04 | 81 | |
| MiniCPM-V 2.0Size=Size < 4B2024.02 | 80.7 | |
| VILA1.5-13BSize=> 8B2024.02 | 80.1 | |
| Bunny-8BSize=4B < Size < 8B2024.02 | 79.9 | |
| Imp-v1.5-4B-Phi3Size=Size < 4B2024.02 | 78.3 | |
| Bunny-4BSize=Size < 4B2024.02 | 78.3 | |
| LLaVA-NeXT-13BSize=> 8B2024.02 | 73.6 | |
| TinyLLaVA-Phi-2-SigLIP-3.1BSize=Size < 4B2024.02 | 73 | |
| MM1-7B-ChatSize=4B < Size < 8B2024.02 | 72.6 | |
| LLaVA-1.5Language Model=Vicuna-13B2023.11 | 71.6 | |
| LLaVA-1.5LLM=Vicuna-13B, PT=558K, IT=665K2024.03 | 71.6 | |
| ShareGPT4V-13BSize=> 8B2024.02 | 71.2 | |
| LLaVA-v1.5-13B (LORA)Size=> 8B2024.02 | 71.2 | |
| Mipha-3BSize=Size < 4B2024.02 | 70.9 | |
| InstructBLIP-Flan-T5-XXLSize=> 8B2024.02 | 70.6 | |
| LVIS-INSTRUCT4V-13BSize=> 8B2024.02 | 70.6 | |
| SPHINX-Intern2Size=4B < Size < 8B2024.02 | 70.4 | |
| LLaVA-NeXT-7BSize=4B < Size < 8B2024.02 | 70.1 | |
| SVIT-v1.5-13BSize=> 8B2024.02 | 70 | |
| MobileVLM-V2-3BSize=Size < 4B2024.02 | 70 | |
| MM1-3B-ChatSize=Size < 4B2024.02 | 69.4 | |
| VILA1.5-3BSize=Size < 4B2024.02 | 69 | |
| mPLUG-Owl2Size=4B < Size < 8B2024.02 | 68.7 | |
| ShareGPT4V-7BLanguage Model=Vicuna-7B2023.11 | 68.4 | |
| LLaVA-PhiLLM=Phi-2-2.7B, PT=558K, IT=665K2024.03 | 68.4 | |
| LLaVA-v1.5-7B (LORA)Size=4B < Size < 8B2024.02 | 68.4 | |
| LLaVA-PhiLLM=Phi 2, Parameters=2.7B2026.04 | 68.4 | |
| Qwen-VL-ChatLanguage Model=Qwen-7B2023.11 | 68.2 | |
| Qwen-VL-ChatLLM=Qwen-7B, PT=1.4B, IT=50M2024.03 | 68.2 | |
| Qwen-VLLanguage Model=Qwen-7B2023.11 | 67.1 | |
| Qwen-VLLLM=Qwen-7B, PT=1.4B, IT=50M2024.03 | 67.1 | |
| LLaVA-1.5Language Model=Vicuna-7B2023.11 | 66.8 | |
| LLaVA-1.5LLM=Vicuna-7B, PT=558K, IT=665K2024.03 | 66.8 | |
| MobileVLM V2LLM=Mobile-LLaMA, Parameters=1.7B2026.04 | 66.7 | |
| VL-MambaLLM=Mamba LLM-2.8B, PT=558K, IT=665K2024.03 | 65.4 | |
| BLIP-2-Flan-T5-XXLSize=> 8B2024.02 | 64.5 | |
| InstructBLIPLanguage Model=FLAN-T52023.11 | 63.1 | |
| InstructBLIPLLM=Vicuna-13B, PT=129M, IT=1.2M2024.03 | 63.1 | |
| InstructBLIP-Vicuna-13BSize=> 8B2024.02 | 63.1 | |
| MoE-LLaVALLM=Qwen, Parameters=2.2B2026.04 | 63.1 | |
| MobileVLM-3BLLM=MobileLLaMA-2.7B, PT=558K, IT=665K2024.03 | 61.2 | |
| BLIP-2Language Model=FLAN-T52023.11 | 61 | |
| BLIP-2LLM=Vicuna-13B, PT=129M2024.03 | 61 | |
| BLIP-2-Vicuna-13BSize=> 8B2024.02 | 61 | |
| InstructBLIPLanguage Model=Vicuna-7B2023.11 | 60.5 | |
| InstructBLIP-Vicuna-7BSize=4B < Size < 8B2024.02 | 60.5 | |
| SmolVLM2LLM=Mobile-LLaMA, Parameters=0.3B2026.04 | 58.8 | |
| MobileVLMLLM=Mobile-LLaMA, Parameters=1.7B2026.04 | 57.3 | |
| Firebolt-VLLLM=LFM2, Parameters=0.8B2026.04 | 56.7 | |
| MiniGPT-4-v2LLM=LLaMA 2, Parameters=8.0B2026.04 | 54.7 | |
| IDEFICSLLM=LLaMA, Parameters=9.0B2026.04 | 53.5 | |
| FastVLLM=Vicuna, Parameters=7.0B2026.04 | 51.1 | |
| InstructBLIPLLM=Vicuna-7B, PT=129M, IT=1.2M2024.03 | 49.2 | |
| ChameleonLLM=LLaMA 2, Parameters=7.0B2026.04 | 46.8 | |
| OpenFlamingo v2LLM=MPT, Parameters=9.0B2026.04 | 44.8 | |
| MiniGPT-4-v1LLM=Vicuna, Parameters=8.0B2026.04 | 39.6 | |
| LLaVALanguage Model=Vicuna-7B2023.11 | 38.5 | |
| Kosmos-2LLM=GPT 2, Parameters=1.7B2026.04 | 32.7 |