EMNLP 2026
The Conference on Empirical Methods in NLP, the most prestigious results driven venue for language model research.
About this event
EMNLP, the Conference on Empirical Methods in Natural Language Processing, is the second of the two flagship NLP venues alongside ACL. Organized since 1996 by SIGDAT, the ACL Special Interest Group on linguistic data and corpus-based approaches, EMNLP emphasizes rigorous empirical methodology and data-driven research. The 2026 edition takes place November 4-9 at the Budapest Congress Center in Hungary, bringing the conference to Central Europe.
The technical program covers the full sweep of modern NLP: large language model pretraining and evaluation, instruction tuning and alignment, retrieval-augmented generation, machine translation, dialogue and agents, low-resource and multilingual languages, and the science of evaluating frontier systems. EMNLP uses the ACL Rolling Review system, with the May 25 ARR deadline feeding into the August acceptance notifications. Acceptance rates hover near 23 percent for the main conference, with additional Findings papers offering a second tier of recognition.
Attendance hovers around 4,500 in person, with the audience tilting toward graduate students, industry research scientists, and faculty from the major NLP labs at Carnegie Mellon, Stanford, Edinburgh, Tsinghua, and the frontier labs. Budapest itself is a draw: a UNESCO heritage city with thermal baths, world class food, and easy rail connections to the rest of Europe. EMNLP closes out the academic calendar each November and is the place to take stock of what the year of LLM progress actually accomplished.