Awards in data, AI and tech

Awards315Companies2,853

315 awards · 2,853 companies

Every award

NeurIPS Best Paper Awards winners by year

NeurIPS · Paper award · usually Nov or Dec · next expected Nov 2026

Official page

Winners

20257 winners

Artificial Hivemind: The Open-Ended Homogeneity of Language Models (and Beyond)Best paperSource ↗
Gated Attention for Large Language Models: Non-linearity, Sparsity, and Attention-Sink-FreeBest paperSource ↗
1000 Layer Networks for Self-Supervised RL: Scaling Depth Can Enable New Goal-Reaching CapabilitiesBest paperSource ↗
Why Diffusion Models Don't Memorize: The Role of Implicit Dynamical Regularization in TrainingBest paperSource ↗
Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?Runner-upSource ↗
Optimal Mistake Bounds for Transductive Online LearningRunner-upSource ↗
Superposition Yields Robust Neural ScalingRunner-upSource ↗

20245 winners

Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale PredictionBest paper, main trackSource ↗
Stochastic Taylor Derivative Estimator: Efficient amortization for arbitrary differential operatorsBest paper, main trackSource ↗
Not All Tokens Are What You Need for PretrainingMain track runner-upSource ↗
Guiding a Diffusion Model with a Bad Version of ItselfMain track runner-upSource ↗
The PRISM Alignment Dataset: What Participatory, Representative and Individualised Human Feedback Reveals About the Subjective and Multicultural Alignment of Large Language ModelsBest paper, datasets and benchmarksSource ↗

20236 winners

Privacy Auditing with One (1) Training RunOutstanding main track paperSource ↗
Are Emergent Abilities of Large Language Models a Mirage?Outstanding main track paperSource ↗
Scaling Data-Constrained Language ModelsMain track runner-upSource ↗
Direct Preference Optimization: Your Language Model is Secretly a Reward ModelMain track runner-upSource ↗
ClimSim: A large multi-scale dataset for hybrid physics-ML climate emulationOutstanding datasets paperSource ↗
DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT ModelsOutstanding benchmarks paperSource ↗
About this award

Best papers and runners-up at the main machine learning conference, main and datasets tracks.

EditionAnnouncedWinnersSource
2025Nov 26, 202511 months ago7
2024Dec 10, 20241 year ago5
2023Dec 11, 20232 years ago6

The next date is an estimate from the months past editions were announced (Nov or Dec); it changes when NeurIPS announces one.

Award detailsSep 25, 2026Now filed as paper award (was research).
Date announcedSep 24, 2026Next edition expected Nov 2026 (was Nov 2025).
New winnersSep 24, 2026Editions 2023 to 2025 loaded with 18 winners; the 2025 edition named 7 papers.