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

Comprehensive Reasoning on Aggregate Benchmarks

66.9Average Accuracy

Qwen2.5-VL-7B + DUPL

Updated 1mo ago

Evaluation Results

Method	Links
Qwen2.5-VL-7B + DUPL 2025.10		66.9
Qwen2.5-VL-7B + NoisyRollout 2025.10		65.4
PAPO-7B 2025.10		65.2
Qwen2.5-VL-7B + GRPO 2025.10		63.9
VLAA-Thinker-7B 2025.10		63.6
MM-Eureka-Qwen-7B 2025.10		63.5
OpenVLThinker-7B 2025.10		61.9
Qwen2.5-VL-3B + DUPL 2025.10		61.2
Qwen2.5-VL-7B 2025.10		61.1
R1-Onevision-7B 2025.10		60.5
Qwen2.5-VL-3B + NoisyRollout 2025.10		59.6
Qwen2.5-VL-3B + GRPO 2025.10		58.4
Qwen2.5-VL-3B 2025.10		52.5