Register and share your invite link to earn from video plays and referrals.

PapersWithCode
@paperswithcodee
Trending AI research papers with code, datasets and leaderboards.
0 Following    234 Followers
Meta FAIR and NYU, with Saining Xie, Lucas Beyer and Luke Zettlemoyer, just released a sharp controlled study on how multimodal models should see images! "Unbundling Multimodal Vision Encoding" Do you need a pretrained vision encoder? They pretrain every model from scratch on the same data at matched compute to find out. It covers why the visual prior isn't necessary, why encoder-free models still trail, how vision modules inside early backbone layers can lower time to first token, and much more! Read this excellent paper by @_ellisbrown and collaborators at @AIatMeta:
Show more
@abel_ai_lab just released Simons, a family of open time series foundation models built with automated research! 📈 "Introducing Simons: Open Time Series Foundation Models" This release gives you zero-shot probabilistic forecasting with the weights, the training code and public training data all out in the open. It covers four models from 25M to 297M parameters, context of up to 32,768 time steps, a 25M model that ranks #3# on the TIME leaderboard on Papers with Code, ahead of Chronos-2 and TimesFM 2.5, and much more! Read more about this release from Abel AI:
Show more
Congrats @sageai_official: Sage S2.1 Flash is now the best open model on OfficeQA Pro! 🏆 It scores 72.2, just 0.7 behind the closed leader, Claude Fable 5.2 (max effort), at 72.9. It also beats Claude Opus 5.5, Sonnet 5.5 and Haiku 5.5, all at max effort. The previous best open model, WorkRover-35B-A3B, scored 48.9.
Show more
Congrats @Kimi_Moonshot: Kimi K3.1 Pro is the best open model on Science's Last Exam! 🏆 It scores 28.92 on 113 real research tasks, ahead of DeepSeek-V4.1-Flash at 26.83. Only GPT-6.1 Sol (44.66) and Claude Opus 5 (37.5) score higher. Can open models do real science yet?
Show more
Yuki M. Asano and collaborators just released a thorough, reproducible recipe for turning pretrained image generators into multi-task perception systems 🍌 "Open Vision Banana: A Fully Open Recipe for Generative Visual Perception" This paper shows you how to frame vision tasks as conditional image editing, and which design choices matter most for generalization. It covers color encoding, prompt design, parameter adaptation and task balancing, 4B and 9B models trained on 2.1M-3.8M pairs from public datasets, 73.9 vs. 69.9 zero-shot mIoU on Cityscapes against Vision Banana, legend ensembling for up to 15.0 points more mIoU on COCO-Object without retraining, and much more! Read this excellent paper by @arpitjadon_ and collaborators:
Show more
@Translation just released Lara Mini base, a small open-weights model trained on English and 44 other languages, covering all 24 official EU languages! 🇪🇺 "Lara Mini Base: Language Diversity as a Resource, Not a Trade-off" This paper shows that in small models, many languages don't have to compete for capacity: multilingual training can help rather than dilute. It covers a 7.2B-total, 0.6B-active MoE trained from scratch on 6T tokens, the highest average score among open base models with up to 2B active parameters (ahead of models trained on 4 to 6x as many tokens), 128k context, generation at least 1.8x faster on one GPU than any higher-scoring model, and much more! Read this excellent paper from the Translated team:
Show more
LFM2.5-24B-A2B from @liquidai scores 96.75 on IFEval. It does that with 2.45B active parameters, out of 24B total. That puts it first on the IFEval instruction-following leaderboard on Papers with Code, ahead of Rufus-Air (95.4) and Qwen3.5-27B (95).
Show more