{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/videoeval-pro-robust-and-realistic-long-video","title":"VideoEval-Pro: Robust and Realistic Long Video Understanding Evaluation","arxiv_id":"2505.14640","date":"2025-05-20","proceeding":null,"authors":["Wentao Ma","Weiming Ren","Yiming Jia","Zhuofeng Li","Ping Nie","Ge Zhang","Wenhu Chen"],"abstract":"Large multimodal models (LMMs) have recently emerged as a powerful tool for long video understanding (LVU), prompting the development of standardized LVU benchmarks to evaluate their performance. However, our investigation reveals a rather sober lesson for existing LVU benchmarks. First, most existing benchmarks rely heavily on multiple-choice questions (MCQs), whose evaluation results are inflated due to the possibility of guessing the correct answer; Second, a significant portion of questions in these benchmarks have strong priors to allow models to answer directly without even reading the input video. For example, Gemini-1.5-Pro can achieve over 50\\% accuracy given a random frame from a long video on Video-MME. We also observe that increasing the number of frames does not necessarily lead to improvement on existing benchmarks, which is counterintuitive. As a result, the validity and robustness of current LVU benchmarks are undermined, impeding a faithful assessment of LMMs' long-video understanding capability. To tackle this problem, we propose VideoEval-Pro, a realistic LVU benchmark containing questions with open-ended short-answer, which truly require understanding the entire video. VideoEval-Pro assesses both segment-level and full-video understanding through perception and reasoning tasks. By evaluating 21 proprietary and open-source video LMMs, we conclude the following findings: (1) video LMMs show drastic performance ($>$25\\%) drops on open-ended questions compared with MCQs; (2) surprisingly, higher MCQ scores do not lead to higher open-ended scores on VideoEval-Pro; (3) compared to other MCQ benchmarks, VideoEval-Pro benefits more from increasing the number of input frames. Our results show that VideoEval-Pro offers a more realistic and reliable measure of long video understanding, providing a clearer view of progress in this domain.","url_abs":"https://arxiv.org/abs/2505.14640v1","url_pdf":"https://arxiv.org/pdf/2505.14640v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"videoeval-pro-robust-and-realistic-long-video","repo_url":"https://github.com/opengvlab/internvideo2","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"videoeval-pro-robust-and-realistic-long-video","repo_url":"https://github.com/opengvlab/videochat-flash","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"mme","task_name":"MME"},{"task_slug":"multiple-choice","task_name":"Multiple-choice"},{"task_slug":null,"task_name":"Video MME"},{"task_slug":"video-understanding","task_name":"Video Understanding"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2505.14640","atlas_url":"https://app.syntology.ai/?focus=2505.14640","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2505.14640"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/opengvlab/videochat-flash","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/opengvlab/internvideo2","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran_fixture":3,"ran_draft_wrong":2,"ran":2,"unverified":7},"by_repo_kind":{"listed":{"samples":14,"ran":7,"repositories":2}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"30d7eec482ebf6b1","entry":"repeat_kv","repo":"opengvlab/videochat-flash","repo_kind":"listed","path":"llava-train_videochat/llava/model/language_model/modeling_qwen2_flash.py","file_url":"https://github.com/opengvlab/videochat-flash/blob/HEAD/llava-train_videochat/llava/model/language_model/modeling_qwen2_flash.py","link_basis":"harvester_set","language":"python","status":"ran_fixture","verification_level":2,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"30d7eec482ebf6b1"}},{"code_sha256_prefix":"d61c483a3c2b3156","entry":"apply_rotary_pos_emb","repo":"opengvlab/videochat-flash","repo_kind":"listed","path":"llava-train_videochat/llava/model/language_model/modeling_qwen2_flash.py","file_url":"https://github.com/opengvlab/videochat-flash/blob/HEAD/llava-train_videochat/llava/model/language_model/modeling_qwen2_flash.py","link_basis":"harvester_set","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"d61c483a3c2b3156"}},{"code_sha256_prefix":"b0a851cae92754ff","entry":"extract_patches","repo":"opengvlab/videochat-flash","repo_kind":"listed","path":"llava-train_videochat/llava/mm_utils.py","file_url":"https://github.com/opengvlab/videochat-flash/blob/HEAD/llava-train_videochat/llava/mm_utils.py","link_basis":"harvester_set","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b0a851cae92754ff"}},{"code_sha256_prefix":"53281c424cf5891f","entry":"find_all_linear_names","repo":"opengvlab/internvideo2","repo_kind":"listed","path":"xtuner-train_internvideo2_5/xtuner/model/utils.py","file_url":"https://github.com/opengvlab/internvideo2/blob/HEAD/xtuner-train_internvideo2_5/xtuner/model/utils.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"53281c424cf5891f"}},{"code_sha256_prefix":"e6bb171e5f95a784","entry":"packing_collate","repo":"opengvlab/internvideo2","repo_kind":"listed","path":"xtuner-train_internvideo2_5/unify_internvl2_train_r16.py","file_url":"https://github.com/opengvlab/internvideo2/blob/HEAD/xtuner-train_internvideo2_5/unify_internvl2_train_r16.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"e6bb171e5f95a784"}},{"code_sha256_prefix":"b99eea6376d1e212","entry":"rotate_half","repo":"opengvlab/videochat-flash","repo_kind":"listed","path":"llava-train_videochat/llava/model/language_model/modeling_qwen2_flash.py","file_url":"https://github.com/opengvlab/videochat-flash/blob/HEAD/llava-train_videochat/llava/model/language_model/modeling_qwen2_flash.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b99eea6376d1e212"}},{"code_sha256_prefix":"2c7b71210b8a363f","entry":"unpad_image","repo":"opengvlab/videochat-flash","repo_kind":"listed","path":"llava-train_videochat/llava/model/llava_arch.py","file_url":"https://github.com/opengvlab/videochat-flash/blob/HEAD/llava-train_videochat/llava/model/llava_arch.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"2c7b71210b8a363f"}},{"code_sha256_prefix":"3c8823948724e269","entry":"auto_pad_images","repo":"opengvlab/videochat-flash","repo_kind":"listed","path":"llava-train_videochat/llava/mm_utils.py","file_url":"https://github.com/opengvlab/videochat-flash/blob/HEAD/llava-train_videochat/llava/mm_utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3c8823948724e269"}},{"code_sha256_prefix":"dcdc7ef233f3e1e8","entry":"build_model","repo":"opengvlab/internvideo2","repo_kind":"listed","path":"xtuner-train_internvideo2_5/xtuner/apis/model.py","file_url":"https://github.com/opengvlab/internvideo2/blob/HEAD/xtuner-train_internvideo2_5/xtuner/apis/model.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"dcdc7ef233f3e1e8"}},{"code_sha256_prefix":"f8035a195671164a","entry":"convert_state_dict_to_hf","repo":"opengvlab/internvideo2","repo_kind":"listed","path":"xtuner-train_internvideo2_5/xtuner/model/llava.py","file_url":"https://github.com/opengvlab/internvideo2/blob/HEAD/xtuner-train_internvideo2_5/xtuner/model/llava.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"f8035a195671164a"}},{"code_sha256_prefix":"269c10dd16c744e8","entry":"create_reference_model","repo":"opengvlab/internvideo2","repo_kind":"listed","path":"xtuner-train_internvideo2_5/xtuner/model/dpo.py","file_url":"https://github.com/opengvlab/internvideo2/blob/HEAD/xtuner-train_internvideo2_5/xtuner/model/dpo.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"269c10dd16c744e8"}},{"code_sha256_prefix":"9356d2ed9b73134b","entry":"get_peft_model_state_dict","repo":"opengvlab/internvideo2","repo_kind":"listed","path":"xtuner-train_internvideo2_5/xtuner/model/utils.py","file_url":"https://github.com/opengvlab/internvideo2/blob/HEAD/xtuner-train_internvideo2_5/xtuner/model/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"9356d2ed9b73134b"}},{"code_sha256_prefix":"a0a407c2a1c1f7d3","entry":"reduce_mean","repo":"opengvlab/internvideo2","repo_kind":"listed","path":"xtuner-train_internvideo2_5/xtuner/model/reward.py","file_url":"https://github.com/opengvlab/internvideo2/blob/HEAD/xtuner-train_internvideo2_5/xtuner/model/reward.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a0a407c2a1c1f7d3"}},{"code_sha256_prefix":"c80358dff96c25ce","entry":"resize_and_center_crop","repo":"opengvlab/videochat-flash","repo_kind":"listed","path":"llava-train_videochat/llava/mm_utils.py","file_url":"https://github.com/opengvlab/videochat-flash/blob/HEAD/llava-train_videochat/llava/mm_utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c80358dff96c25ce"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}