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

Hallucination Evaluation on HalBench

76.7Score

Qwen3-VL-8B-Thinking + DiG

Updated 25d ago

Evaluation Results

Method	Links
Qwen3-VL-8B-Thinking + DiG 2025.12		76.7
Qwen3-VL-4B-Thinking + DiG 2025.12		73.8
Qwen3-VL-8B-Thinking 2025.12		73.3
Qwen3-VL-4B-Thinking 2025.12		70.1
DAMO 2026.06		58.3
PAI 2026.06		57.8
FADE 2026.06		57.2
InternVL2.5-78B 2025.12		57.1
Greedy 2026.06		56.8
VCD 2026.06		56.2
DCLA 2026.06		56.1
Greedy 2026.06		55.9
PAI 2026.06		55.7
VCD 2026.06		55.3
Qwen2.5-VL-72B 2025.12		55.2
DAMO 2026.06		55.2
FADE 2026.06		55.2
DCLA 2026.06		54.9
DAMO 2026.06		51.94
Greedy 2026.06		50.32
Qwen2.5-VL-7B 2025.12		50.1
PAI 2026.06		50.06
DCLA 2026.06		49.23
FADE 2026.06		48.34
VCD 2026.06		47.5