Share your thoughts, 1 month free Claude Pro on usSee more

Multimodal Reasoning on SEED-BENCH

69.9Accuracy

InternVL2-8B + RP

Updated 1d ago

Evaluation Results

Method	Links
InternVL2-8B + RP 2024.08		69.9
InternVL2-8B-FT 2024.08		69.5
MGM-7B 2024.08		63.5
MGM-7B + RP 2024.08		63.2
LLaVA-1.5-13B + MMFuser 2024.10		62
LLaVA-1.5-7B + RP 2024.08		61.7
LLaVA-1.5-13B 2024.10		61.6
LLaVA-1.5-7B + MMFuser 2024.10		60.8
Vanilla 2026.06		59.3
LLaVA-1.5-7B 2024.10		58.6
LLaVA-1.5-7B 2024.08		58.6
Qwen-VL-Chat 2024.10		58.2
EvoCut 2026.06		57.3
FastV 2026.06		57.1
ApET 2026.06		56.8
VisionZip 2026.06		56.4
V2Drop 2026.06		56.4
Qwen-VL 2024.10		56.3
FastV 2026.06		55.9
SparseVLM 2026.06		55.8
EvoCut 2026.06		55.2
VisionZip 2026.06		54.9
PDrop 2026.06		54.7
ApET 2026.06		54.7
V2Drop 2026.06		53.8
InstructBLIP 2024.10		53.4
SparseVLM 2026.06		53.4
PDrop 2026.06		53.3
EvoCut 2026.06		53.2
ApET 2026.06		52.4
VisionZip 2026.06		52.2
FastV 2026.06		51.9
V2Drop 2026.06		51.4
SparseVLM 2026.06		51.1
BLIP-2 2024.10		46.4
PDrop 2026.06		40