{"url":"/sota/question-answering-on-newsqa","task":{"name":"Question Answering","url":"/task/question-answering","note":null},"dataset":{"name":"NewsQA","url":"/dataset/newsqa"},"category":"Natural Language Processing","categories":["Miscellaneous","Natural Language Processing","Reasoning"],"category_note":null,"description":"Question answering can be segmented into domain-specific tasks like community question answering and knowledge-base question answering. Popular benchmark datasets for evaluation question answering systems include [SQuAD](/dataset/squad), [HotPotQA](/dataset/hotpotqa), [bAbI](/dataset/babi-1), [TriviaQA](/dataset/triviaqa), [WikiQA](/dataset/wikiqa), and many others. Models for question answering are typically evaluated on metrics like EM and F1. Some recent top performing models are T5 and XLNet.\r\n\r\n( Image credit: [SQuAD](https://rajpurkar.github.io/mlx/qa-and-squad/) )","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":["EM","F1"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"EM":null,"F1":"higher"}},"counts":{"rows":18,"rows_with_code":14,"rows_with_paper_page":18,"rows_dated":17,"rows_using_additional_data":12},"rows":[{"rank_in_archive_order":1,"model":"OpenAI/o3-2025-01-31-high","metrics":{"EM":"92.52","F1":"93.13"},"uses_additional_data":true,"paper_date":"2025-01-30","paper":"/paper/o3-mini-vs-deepseek-r1-which-one-is-safer","paper_url":"https://arxiv.org/abs/2501.18438v2","paper_title":"o3-mini vs DeepSeek-R1: Which One is Safer?","code":"https://github.com/trust4ai/astral","n_code_links":1,"syntology":null},{"rank_in_archive_order":2,"model":"Riple/Saanvi-v0.5-DeepAnalysis","metrics":{"EM":"92.14","F1":"94.01"},"uses_additional_data":true,"paper_date":"2016-11-07","paper":"/paper/deepsense-a-unified-deep-learning-framework","paper_url":"http://arxiv.org/abs/1611.01942v2","paper_title":"DeepSense: A Unified Deep Learning Framework for Time-Series Mobile Sensing Data Processing","code":"https://github.com/DigitalBiomarkerDiscoveryPipeline/Human-Activity-Recognition","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"OpenAI/o4-mini-2025-05-01-high","metrics":{"EM":"88.24","F1":"91.31"},"uses_additional_data":true,"paper_date":"2021-06-13","paper":"/paper/thinking-like-transformers-1","paper_url":"https://arxiv.org/abs/2106.06981v2","paper_title":"Thinking Like Transformers","code":"https://github.com/google-deepmind/tracr","n_code_links":5,"syntology":{"n_ran":11,"n_unverified":7,"n_samples":18,"n_pointer_only_licence":5}},{"rank_in_archive_order":4,"model":"OpenAI/o1-2024-12-17-high","metrics":{"EM":"81.44","F1":"88.72"},"uses_additional_data":true,"paper_date":"2022-06-19","paper":"/paper/0-1-deep-neural-networks-via-block-coordinate","paper_url":"https://arxiv.org/abs/2206.09379v2","paper_title":"0/1 Deep Neural Networks via Block Coordinate Descent","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":5,"model":"deepseek-r1","metrics":{"EM":"80.57","F1":"86.13"},"uses_additional_data":true,"paper_date":"2025-01-22","paper":"/paper/deepseek-r1-incentivizing-reasoning","paper_url":"https://arxiv.org/abs/2501.12948v1","paper_title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","code":"https://github.com/deepseek-ai/deepseek-r1","n_code_links":4,"syntology":null},{"rank_in_archive_order":6,"model":"Anthropic/claude-3-7-sonnet","metrics":{"EM":"74.23","F1":"82.3"},"uses_additional_data":false,"paper_date":"2024-06-24","paper":"/paper/claude-3-5-sonnet-model-card-addendum","paper_url":"https://www-cdn.anthropic.com/fed9cc193a14b84131812372d8d5857f8f304c52/Model_Card_Claude_3_Addendum.pdf","paper_title":"Claude 3.5 Sonnet Model Card Addendum","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":7,"model":"Riple/Saanvi-v0.1","metrics":{"EM":"72.61","F1":"85.44"},"uses_additional_data":false,"paper_date":"2022-05-23","paper":"/paper/time-series-transformer-generative","paper_url":"https://arxiv.org/abs/2205.11164v1","paper_title":"Time-series Transformer Generative Adversarial Networks","code":"https://github.com/jsyoon0823/TimeGAN","n_code_links":5,"syntology":null},{"rank_in_archive_order":8,"model":"xAI/grok-3-1212","metrics":{"EM":"70.57","F1":"88.24"},"uses_additional_data":true,"paper_date":"2022-02-15","paper":"/paper/xai-for-transformers-better-explanations","paper_url":"https://arxiv.org/abs/2202.07304v2","paper_title":"XAI for Transformers: Better Explanations through Conservative Propagation","code":"https://github.com/ameenali/xai_transformers","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":10,"n_samples":12,"n_pointer_only_licence":0}},{"rank_in_archive_order":9,"model":"OpenAI/GPT-4o","metrics":{"EM":"70.21","F1":"81.74"},"uses_additional_data":true,"paper_date":"2024-10-03","paper":"/paper/sieve-general-purpose-data-filtering-system","paper_url":"https://arxiv.org/abs/2410.02755v3","paper_title":"GPT-4o as the Gold Standard: A Scalable and General Purpose Approach to Filter Language Model Pretraining Data","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":10,"model":"Google/Gemini 2.5 Pro","metrics":{"EM":"68.75","F1":"79.91"},"uses_additional_data":true,"paper_date":"2024-03-08","paper":"/paper/gemini-1-5-unlocking-multimodal-understanding","paper_url":"https://arxiv.org/abs/2403.05530v5","paper_title":"Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context","code":"https://github.com/dlvuldet/primevul","n_code_links":1,"syntology":null},{"rank_in_archive_order":11,"model":"BERT+ASGen","metrics":{"EM":"54.7","F1":"64.5"},"uses_additional_data":false,"paper_date":null,"paper":"/paper/learning-to-generate-questions-by-learning-to","paper_url":"https://aclanthology.org/2021.findings-acl.132","paper_title":"Learning to Generate Questions by Learning to Recover Answer-containing Sentences","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":12,"model":"DecaProp","metrics":{"EM":"53.1","F1":"66.3"},"uses_additional_data":false,"paper_date":"2018-11-10","paper":"/paper/densely-connected-attention-propagation-for","paper_url":"http://arxiv.org/abs/1811.04210v2","paper_title":"Densely Connected Attention Propagation for Reading Comprehension","code":"https://github.com/vanzytay/NIPS2018_DECAPROP","n_code_links":2,"syntology":null},{"rank_in_archive_order":13,"model":"MINIMAL(Dyn)","metrics":{"EM":"50.1","F1":"63.2"},"uses_additional_data":true,"paper_date":"2018-05-21","paper":"/paper/efficient-and-robust-question-answering-from","paper_url":"http://arxiv.org/abs/1805.08092v1","paper_title":"Efficient and Robust Question Answering from Minimal Context over Documents","code":"https://github.com/SatyamSoni23/Smart-Question-Answering-System-on-Document","n_code_links":1,"syntology":null},{"rank_in_archive_order":14,"model":"AMANDA","metrics":{"EM":"48.4","F1":"63.7"},"uses_additional_data":false,"paper_date":"2018-01-25","paper":"/paper/a-question-focused-multi-factor-attention","paper_url":"http://arxiv.org/abs/1801.08290v1","paper_title":"A Question-Focused Multi-Factor Attention Network for Question Answering","code":"https://github.com/nusnlp/amanda","n_code_links":1,"syntology":null},{"rank_in_archive_order":15,"model":"FastQAExt","metrics":{"EM":"43.7","F1":"56.1"},"uses_additional_data":true,"paper_date":"2017-03-14","paper":"/paper/making-neural-qa-as-simple-as-possible-but","paper_url":"http://arxiv.org/abs/1703.04816v3","paper_title":"Making Neural QA as Simple as Possible but not Simpler","code":"https://github.com/uclmr/jack","n_code_links":3,"syntology":{"n_ran":0,"n_unverified":2,"n_samples":2,"n_pointer_only_licence":0}},{"rank_in_archive_order":16,"model":"SpanBERT","metrics":{"F1":"73.6"},"uses_additional_data":false,"paper_date":"2019-07-24","paper":"/paper/spanbert-improving-pre-training-by","paper_url":"https://arxiv.org/abs/1907.10529v3","paper_title":"SpanBERT: Improving Pre-training by Representing and Predicting Spans","code":"https://github.com/facebookresearch/SpanBERT","n_code_links":6,"syntology":{"n_ran":3,"n_unverified":12,"n_samples":15,"n_pointer_only_licence":4}},{"rank_in_archive_order":17,"model":"LinkBERT (large)","metrics":{"F1":"72.6"},"uses_additional_data":true,"paper_date":"2022-03-29","paper":"/paper/linkbert-pretraining-language-models-with","paper_url":"https://arxiv.org/abs/2203.15827v1","paper_title":"LinkBERT: Pretraining Language Models with Document Links","code":"https://github.com/michiyasunaga/LinkBERT","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":14,"n_samples":14,"n_pointer_only_licence":0}},{"rank_in_archive_order":18,"model":"DyREX","metrics":{"F1":"68.53"},"uses_additional_data":true,"paper_date":"2022-10-26","paper":"/paper/dyrex-dynamic-query-representation-for","paper_url":"https://arxiv.org/abs/2210.15048v1","paper_title":"DyREx: Dynamic Query Representation for Extractive Question Answering","code":"https://github.com/urchade/dyrex","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":3}}],"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 6,795 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":6795,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2785},"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":6,"rows_with_any_sample_ran":4,"distinct_papers_with_graph_line":6,"distinct_papers_with_any_sample_ran":4,"samples_over_distinct_papers":{"n_ran":18,"n_unverified":46,"n_samples":64,"n_pointer_only_licence":12,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":18,"n_unverified":46,"n_samples":64,"n_pointer_only_licence":12,"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"}}}