{"benchmark_id":"pope","benchmark_name":"POPE","benchmark_description":"Polling-based Object Probing Evaluation (POPE) is a benchmark for evaluating object hallucination in Large Vision-Language Models (LVLMs). POPE addresses the problem where LVLMs generate objects inconsistent with target images by using a polling-based query method that asks yes/no questions about object presence in images, providing more stable and flexible evaluation of object hallucination.","max_score":1.0,"categories":["multimodal","safety","vision"],"modality":"multimodal","total_models":3,"entries":[{"rank":1,"model_id":"lfm-2.5-vl-3b","model_name":"LFM2.5-VL-3B","organization_name":"Liquid AI","organization_id":"liquid","benchmark_score":0.887,"normalized_score":0.887,"verified":false,"self_reported":true,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"speed_rps":null,"context_window":null,"release_date":"2026-08-12","announcement_date":"2026-08-12","multimodal":true,"param_count":3123483888,"is_new":false},{"rank":2,"model_id":"phi-3.5-vision-instruct","model_name":"Phi-3.5-vision-instruct","organization_name":"Microsoft","organization_id":"microsoft","benchmark_score":0.861,"normalized_score":0.861,"verified":false,"self_reported":true,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"speed_rps":null,"context_window":null,"release_date":"2024-08-23","announcement_date":"2024-08-23","multimodal":true,"param_count":4200000000,"is_new":false},{"rank":3,"model_id":"phi-4-multimodal-instruct","model_name":"Phi-4-multimodal-instruct","organization_name":"Microsoft","organization_id":"microsoft","benchmark_score":0.856,"normalized_score":0.856,"verified":false,"self_reported":true,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"speed_rps":null,"context_window":null,"release_date":"2025-02-01","announcement_date":"2025-02-01","multimodal":true,"param_count":5600000000,"is_new":false}]}