List your company
ResearchFormerly WizWand

General VQA on HallusionBench

73.48AccuracyGemini 3-Pro
40.813649.294357.77566.2557Feb 4, 2026Feb 21, 2026Mar 11, 2026Mar 29, 2026Apr 15, 2026May 3, 2026May 21, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Gemini 3-Pro2026.0273.48
Qwen2.5-VL-7B-Instruct + VPPOBackbone model=Qwen2.5-VL-7B-Instruct, Training Data=19.2K2026.0570.8
Perception-R1-7BTraining Data=1.4K2026.0570
Qwen2.5-VL-7B-Instruct + Faithful-MR1Backbone model=Qwen2.5-VL-7B-Instruct, Training Data=19.2K2026.0569.8
Qwen2.5-VL-7B-Instruct + GRPOBackbone model=Qwen2.5-VL-7B-Instruct, Training Data=19.2K2026.0569.3
Vision-R1-7BTraining Data=210K2026.0568.1
Vision-SR1-7BTraining Data=56K2026.0568.1
Qwen2.5-VL-3B-Instruct + Faithful-MR1Backbone model=Qwen2.5-VL-3B-Instruct, Training Data=19.2K2026.0568
Qwen2.5-VL-3B-Instruct + VPPOBackbone model=Qwen2.5-VL-3B-Instruct, Training Data=19.2K2026.0567.7
Qwen2.5-VL-3B-Instruct + GRPOBackbone model=Qwen2.5-VL-3B-Instruct, Training Data=19.2K2026.0567.2
GPT-5tier=High2026.0266.58
Qwen2.5-VL-7B-InstructBackbone model=Qwen2.5-VL-7B-Instruct2026.0565
Qwen2.5-VL-3B-InstructBackbone model=Qwen2.5-VL-3B-Instruct2026.0564.8
Qwen3-VLmode=Thinking2026.0264.01
ERNIE 5.02026.0263.87
Gemini 2.5-Pro2026.0263.7
Qwen3-VLLanguage Backbone=Qwen3 1.7B LLM, Evaluation Framework=VLMEvalKit [69]2026.0351.89
Intern3.5-VLLanguage Backbone=Qwen3 1.7B LLM, Evaluation Framework=VLMEvalKit [69]2026.0348.18
FineViT-VLLanguage Backbone=Qwen3 1.7B LLM, Evaluation Framework=VLMEvalKit [69]2026.0346.54
Aquila-VLLanguage Backbone=Qwen2.5 1.5B LLM, Evaluation Framework=VLMEvalKit [69]2026.0342.07