{"url":"/sota/visual-question-answering-vqa-on","task":{"name":"Visual Question Answering (VQA)","url":"/task/visual-question-answering","note":null},"dataset":{"name":"InfographicVQA","url":"/dataset/infographicvqa"},"category":"Computer Vision","categories":["Computer Vision","Natural Language Processing"],"category_note":null,"description":"**Visual Question Answering (VQA)** is a task in computer vision that involves answering questions about an image. The goal of VQA is to teach machines to understand the content of an image and answer questions about it in natural language.\r\n\r\nImage Source: [visualqa.org](https://visualqa.org/)","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["ANLS"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"ANLS":null}},"counts":{"rows":21,"rows_with_code":16,"rows_with_paper_page":21,"rows_dated":21,"rows_using_additional_data":11},"rows":[{"rank_in_archive_order":1,"model":"Gemini Ultra (pixel only)","metrics":{"ANLS":"80.3"},"uses_additional_data":false,"paper_date":"2023-12-19","paper":"/paper/gemini-a-family-of-highly-capable-multimodal-1","paper_url":"https://arxiv.org/abs/2312.11805v5","paper_title":"Gemini: A Family of Highly Capable Multimodal Models","code":"https://github.com/valdecy/pybibx","n_code_links":1,"syntology":null},{"rank_in_archive_order":2,"model":"SMoLA-PaLI-X Specialist","metrics":{"ANLS":"66.2"},"uses_additional_data":true,"paper_date":"2023-12-01","paper":"/paper/omni-smola-boosting-generalist-multimodal","paper_url":"https://arxiv.org/abs/2312.00968v2","paper_title":"Omni-SMoLA: Boosting Generalist Multimodal Models with Soft Mixture of Low-rank Experts","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":3,"model":"ScreenAI 5B (4.62 B params, w/ OCR)","metrics":{"ANLS":"65.90"},"uses_additional_data":true,"paper_date":"2024-02-07","paper":"/paper/screenai-a-vision-language-model-for-ui-and","paper_url":"https://arxiv.org/abs/2402.04615v3","paper_title":"ScreenAI: A Vision-Language Model for UI and Infographics Understanding","code":"https://github.com/google-research-datasets/screen_qa","n_code_links":2,"syntology":null},{"rank_in_archive_order":4,"model":"SMoLA-PaLI-X Generalist","metrics":{"ANLS":"65.6"},"uses_additional_data":true,"paper_date":"2023-12-01","paper":"/paper/omni-smola-boosting-generalist-multimodal","paper_url":"https://arxiv.org/abs/2312.00968v2","paper_title":"Omni-SMoLA: Boosting Generalist Multimodal Models with Soft Mixture of Low-rank Experts","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":5,"model":"UDOP (aux)","metrics":{"ANLS":"63.0"},"uses_additional_data":true,"paper_date":"2022-12-05","paper":"/paper/unifying-vision-text-and-layout-for-universal","paper_url":"https://arxiv.org/abs/2212.02623v3","paper_title":"Unifying Vision, Text, and Layout for Universal Document Processing","code":"https://github.com/microsoft/i-code","n_code_links":5,"syntology":{"n_ran":4,"n_unverified":13,"n_samples":17,"n_pointer_only_licence":3}},{"rank_in_archive_order":6,"model":"PaLI-3 (w/ OCR)","metrics":{"ANLS":"62.4"},"uses_additional_data":false,"paper_date":"2023-10-13","paper":"/paper/pali-3-vision-language-models-smaller-faster","paper_url":"https://arxiv.org/abs/2310.09199v2","paper_title":"PaLI-3 Vision Language Models: Smaller, Faster, Stronger","code":"https://github.com/kyegomez/PALI3","n_code_links":1,"syntology":{"n_ran":3,"n_unverified":0,"n_samples":3,"n_pointer_only_licence":0}},{"rank_in_archive_order":7,"model":"TILT-Large","metrics":{"ANLS":"61.20"},"uses_additional_data":true,"paper_date":"2021-02-18","paper":"/paper/going-full-tilt-boogie-on-document","paper_url":"https://arxiv.org/abs/2102.09550v3","paper_title":"Going Full-TILT Boogie on Document Understanding with Text-Image-Layout Transformer","code":"https://github.com/uakarsh/TiLT-Implementation","n_code_links":1,"syntology":null},{"rank_in_archive_order":8,"model":"PaLI-3","metrics":{"ANLS":"57.8"},"uses_additional_data":false,"paper_date":"2023-10-13","paper":"/paper/pali-3-vision-language-models-smaller-faster","paper_url":"https://arxiv.org/abs/2310.09199v2","paper_title":"PaLI-3 Vision Language Models: Smaller, Faster, Stronger","code":"https://github.com/kyegomez/PALI3","n_code_links":1,"syntology":{"n_ran":3,"n_unverified":0,"n_samples":3,"n_pointer_only_licence":0}},{"rank_in_archive_order":9,"model":"ChatGPT 3.5 with LAPDoc Prompt (SpatialFormat)","metrics":{"ANLS":"54.9"},"uses_additional_data":false,"paper_date":"2024-02-15","paper":"/paper/lapdoc-layout-aware-prompting-for-documents","paper_url":"https://arxiv.org/abs/2402.09841v1","paper_title":"LAPDoc: Layout-Aware Prompting for Documents","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":10,"model":"PaLI-X (Single-task FT w/ OCR)","metrics":{"ANLS":"54.8"},"uses_additional_data":true,"paper_date":"2023-05-29","paper":"/paper/pali-x-on-scaling-up-a-multilingual-vision","paper_url":"https://arxiv.org/abs/2305.18565v1","paper_title":"PaLI-X: On Scaling up a Multilingual Vision and Language Model","code":"https://github.com/kyegomez/PALI","n_code_links":2,"syntology":{"n_ran":6,"n_unverified":1,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":11,"model":"Claude + LATIN-Prompt","metrics":{"ANLS":"54.51"},"uses_additional_data":false,"paper_date":"2023-06-01","paper":"/paper/layout-and-task-aware-instruction-prompt-for","paper_url":"https://arxiv.org/abs/2306.00526v4","paper_title":"Layout and Task Aware Instruction Prompt for Zero-shot Document Image Question Answering","code":"https://github.com/wenjinw/latin-prompt","n_code_links":3,"syntology":{"n_ran":1,"n_unverified":9,"n_samples":10,"n_pointer_only_licence":0}},{"rank_in_archive_order":12,"model":"PaLI-X (Multi-task FT)","metrics":{"ANLS":"50.7"},"uses_additional_data":true,"paper_date":"2023-05-29","paper":"/paper/pali-x-on-scaling-up-a-multilingual-vision","paper_url":"https://arxiv.org/abs/2305.18565v1","paper_title":"PaLI-X: On Scaling up a Multilingual Vision and Language Model","code":"https://github.com/kyegomez/PALI","n_code_links":2,"syntology":{"n_ran":6,"n_unverified":1,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":13,"model":"PaLI-X (Single-task FT)","metrics":{"ANLS":"49.2"},"uses_additional_data":true,"paper_date":"2023-05-29","paper":"/paper/pali-x-on-scaling-up-a-multilingual-vision","paper_url":"https://arxiv.org/abs/2305.18565v1","paper_title":"PaLI-X: On Scaling up a Multilingual Vision and Language Model","code":"https://github.com/kyegomez/PALI","n_code_links":2,"syntology":{"n_ran":6,"n_unverified":1,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":14,"model":"GPT-3.5 + LATIN-Prompt","metrics":{"ANLS":"48.98"},"uses_additional_data":false,"paper_date":"2023-06-01","paper":"/paper/layout-and-task-aware-instruction-prompt-for","paper_url":"https://arxiv.org/abs/2306.00526v4","paper_title":"Layout and Task Aware Instruction Prompt for Zero-shot Document Image Question Answering","code":"https://github.com/wenjinw/latin-prompt","n_code_links":3,"syntology":{"n_ran":1,"n_unverified":9,"n_samples":10,"n_pointer_only_licence":0}},{"rank_in_archive_order":15,"model":"DocFormerv2-large","metrics":{"ANLS":"48.8"},"uses_additional_data":true,"paper_date":"2023-06-02","paper":"/paper/docformerv2-local-features-for-document","paper_url":"https://arxiv.org/abs/2306.01733v1","paper_title":"DocFormerv2: Local Features for Document Understanding","code":"https://github.com/uakarsh/docformerv2","n_code_links":1,"syntology":null},{"rank_in_archive_order":16,"model":"UDOP","metrics":{"ANLS":"47.4"},"uses_additional_data":false,"paper_date":"2022-12-05","paper":"/paper/unifying-vision-text-and-layout-for-universal","paper_url":"https://arxiv.org/abs/2212.02623v3","paper_title":"Unifying Vision, Text, and Layout for Universal Document Processing","code":"https://github.com/microsoft/i-code","n_code_links":5,"syntology":{"n_ran":4,"n_unverified":13,"n_samples":17,"n_pointer_only_licence":3}},{"rank_in_archive_order":17,"model":"DUBLIN (variable resolution)","metrics":{"ANLS":"42.6"},"uses_additional_data":true,"paper_date":"2023-05-23","paper":"/paper/dublin-document-understanding-by-language","paper_url":"https://arxiv.org/abs/2305.14218v4","paper_title":"DUBLIN -- Document Understanding By Language-Image Network","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":18,"model":"Pix2Struct-large","metrics":{"ANLS":"40"},"uses_additional_data":false,"paper_date":"2022-10-07","paper":"/paper/pix2struct-screenshot-parsing-as-pretraining","paper_url":"https://arxiv.org/abs/2210.03347v2","paper_title":"Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understanding","code":"https://github.com/huggingface/transformers","n_code_links":4,"syntology":{"n_ran":0,"n_unverified":5,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":19,"model":"Pix2Struct-base","metrics":{"ANLS":"38.2"},"uses_additional_data":false,"paper_date":"2022-10-07","paper":"/paper/pix2struct-screenshot-parsing-as-pretraining","paper_url":"https://arxiv.org/abs/2210.03347v2","paper_title":"Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understanding","code":"https://github.com/huggingface/transformers","n_code_links":4,"syntology":{"n_ran":0,"n_unverified":5,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":20,"model":"MatCha","metrics":{"ANLS":"37.2"},"uses_additional_data":false,"paper_date":"2022-12-19","paper":"/paper/matcha-enhancing-visual-language-pretraining","paper_url":"https://arxiv.org/abs/2212.09662v2","paper_title":"MatCha: Enhancing Visual Language Pretraining with Math Reasoning and Chart Derendering","code":"https://github.com/huggingface/transformers","n_code_links":1,"syntology":null},{"rank_in_archive_order":21,"model":"DUBLIN","metrics":{"ANLS":"36.82"},"uses_additional_data":true,"paper_date":"2023-05-23","paper":"/paper/dublin-document-understanding-by-language","paper_url":"https://arxiv.org/abs/2305.14218v4","paper_title":"DUBLIN -- Document Understanding By Language-Image Network","code":null,"n_code_links":0,"syntology":null}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 6,264 of the 9,581 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9581,"papers_checked":6264,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":3316},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":11,"rows_with_any_sample_ran":9,"distinct_papers_with_graph_line":5,"distinct_papers_with_any_sample_ran":4,"samples_over_distinct_papers":{"n_ran":14,"n_unverified":28,"n_samples":42,"n_pointer_only_licence":3,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":34,"n_unverified":57,"n_samples":91,"n_pointer_only_licence":6,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}