Releases, commentary, and research updates from LAION.https://laion.ai/enhttps://laion.ai/blog/bvd/https://laion.ai/blog/bvd/Today we are releasing LAION-BVD (Big Video Dataset), a large-scale open video dataset designed for multimodal machine learning research. LAION-BVD contains: 1.3 billion platform-specific video URLs collected from Common Crawl 80 million downloaded videos 10 million hours of video content 55 million captioned video cli…Wed, 26 Aug 2026 00:00:00 GMTAndreas Hochlehnert, Marianna Nezhurina, Mehdi Cherti, Andrej Radonjic, Thaddäus Wiedemer, Christoph Schuhmann, Romain Beaumont, Wieland Brendel, Bernhard Schölkopf, A. Sophia Koepke, Jenia Jitsev, Matthias Bethgehttps://laion.ai/blog/open-sci-ref-001/https://laion.ai/blog/open-sci-ref-001/We introduce and release open-sci-ref 0.01 - a research dense transformer model family with all the intermediate checkpoints trained on 8 different well established reference datasets on various model (0.13B - 0.4B - 1.3B - 1.7B) and token (50B, 300B, 1T) scales to serve as baselines for comparison and for studies on t…Mon, 18 Aug 2025 00:00:00 GMTMarianna Nezhurina, Joerg Franke, David Salinas, Jenia Jitsev; open-sci collective, openEuroLLM team, LAIONhttps://laion.ai/blog/reasoning_game_arena_blog_post/https://laion.ai/blog/reasoning_game_arena_blog_post/Access - Repository : https://github.com/SLAMPAI/game reasoning arena - Documentation : Complete installation, usage, and extension guides available at Game Reasoning Arena Documentation - Paper : Game Reasoning Arena: A Framework and Benchmark for Assessing Reasoning Capabilites of Large Language Models via Game Play…Mon, 04 Aug 2025 00:00:00 GMTLucia Cipolina-Kun, Marianna Nezhurina, Jenia Jitsevhttps://laion.ai/blog/do-they-see-what-we-see/https://laion.ai/blog/do-they-see-what-we-see/Building Emotionally Intelligent AI with EmoNet, a suite of open tools and resources developed by the LAION community A LAION & Intel Collaboration Authors LAION e.V. Christoph Schuhmann , Robert Kaczmarczyk , Gollam Rabby, Felix Friedrich, Maurice Kraus, Kourosh Nadi, Desmond Grealy, Huu Nguyen, Cahya Wirawan, Krishna…Thu, 19 Jun 2025 00:00:00 GMTLAIONhttps://laion.ai/blog/bud-e-release/https://laion.ai/blog/bud-e-release/Today marks a milestone in our journey towards democratizing education and empathy through technology. LAION e.V. is thrilled to announce the release of BUD-E version 1.0, an open-source, privacy-compliant AI education assistant framework. BUD-E, which stands for Buddy for Understanding and Digital Empathy, represents…Mon, 20 Jan 2025 00:00:00 GMTLAIONhttps://laion.ai/blog/laion-disco-12m/https://laion.ai/blog/laion-disco-12m/LAION announces the LAION-DISCO-12M - a collection of 12 million links to publicly available YouTube samples paired with metadata to support basic machine learning research in foundation models for generic audio, music information retrieval, and audio dataset analysis. We collect metadata for 12,648,485 songs, includin…Sun, 17 Nov 2024 00:00:00 GMTLAION e.V.https://laion.ai/blog/laion-intel-cooperation/https://laion.ai/blog/laion-intel-cooperation/Large-scale Artificial Intelligence Open Network (LAION) is honored to become an Intel AI/oneAPI Center of Excellence to advance the development of BUD-E (Buddy for Understanding and Digital Empathy) , an open source, empathetic AI education assistant aimed at democratizing personalized learning worldwide. Leading the…Tue, 10 Sep 2024 00:00:00 GMTLAION e.V.https://laion.ai/blog/relaion-5b/https://laion.ai/blog/relaion-5b/Today, following a safety revision procedure, we announce Re-LAION-5B, an updated version of LAION-5B, that is the first web-scale, text-link to images pair dataset to be thoroughly cleaned of known links to suspected CSAM. Highlights - Re-LAION-5B fixes the issues as reported by Stanford Internet Observatory in Decemb…Fri, 30 Aug 2024 00:00:00 GMTLAION e.V.https://laion.ai/blog/bud-e/https://laion.ai/blog/bud-e/AI voice assistants have revolutionized our interaction with technology, answering queries, performing tasks, and making life easier. However, the stilted, mechanical nature of their responses is a barrier to truly immersive conversational experiences. Unlike human conversation partners, they often struggle with fully…Thu, 08 Feb 2024 00:00:00 GMTLAIONhttps://laion.ai/blog/laion-pop/https://laion.ai/blog/laion-pop/LAION POP is a subset of LAION 5B: This subset comprises 600,000 high-resolution images, each equipped with detailed descriptions. The selection of images was based on 10,000 different concepts popular on the image generation site "Midjourney". SampleOverview :------------------------------: LAION-POP Dataset on Huggin…Fri, 17 Nov 2023 00:00:00 GMTChristoph Schuhmann, Peter Bevanhttps://laion.ai/blog/open-empathic/https://laion.ai/blog/open-empathic/We are thrilled to present Open Empathic , a pioneering open-source project initiated by our non-profit organization, LAION. Open Empathic aims to equip open-source AI systems with empathy and emotional intelligence. We hope that methods and tools developed within the framework of this project, together with a communit…Sun, 22 Oct 2023 00:00:00 GMTChristoph, Knoriy, Roberthttps://laion.ai/blog/strategic-game-dataset/https://laion.ai/blog/strategic-game-dataset/Recent advancements in artificial intelligence (AI) underscore the progress of reasoning and planning shown by recent generalist machine learning (ML) models. The progress can be boosted by datasets that can further boost these generic capabilities when used for training foundation models of various kind. This research…Wed, 18 Oct 2023 00:00:00 GMTChristoph Schuhmann & Qi Sunhttps://laion.ai/blog/clara-release/https://laion.ai/blog/clara-release/Voices carry not only words but also convey emotions, emphasis, and nuance through aspects like tone and accent. However, existing speech technology only partially comprehends these intricate components of human speech. Introducing CLARA (Multilingual Contrastive Learning for Audio Representation Acquisition) – a proje…Mon, 16 Oct 2023 00:00:00 GMTKnoriy, Christoph, Roberthttps://laion.ai/blog/leo-lm/https://laion.ai/blog/leo-lm/We proudly introduce LeoLM ( L inguistically E nhanced O pen L anguage M odel), the first comprehensive suite of German-language Foundation Language Models trained in collaboration with HessianAI on their new supercomputer 42 ! Built on Llama-2 and trained on a large-scale, high-quality German text corpus, we present L…Thu, 28 Sep 2023 00:00:00 GMTBjörn Plüsterhttps://laion.ai/blog/open-lm/https://laion.ai/blog/open-lm/Introduction We release OpenLM a simple and minimalist PyTorch codebase for training medium-sized language models. OpenLM is designed to maximize GPU utilization and training speed, and is easy to modify for new language model research and applications. We validate OpenLM by training two language models, OpenLM-1B and…Tue, 26 Sep 2023 00:00:00 GMTOpenLM teamhttps://laion.ai/blog/transparent-ai/https://laion.ai/blog/transparent-ai/Following our previous open letter to the European Parliament on the significance of open-source AI, LAION, backed by European Laboratory for Learning and Intelligent Systems (ELLIS) and a long list of very impactful AI researchers, we submit this new open letter to the European Parliament: Link to the PDF ---------- W…Thu, 21 Sep 2023 00:00:00 GMTLAIONhttps://laion.ai/blog/falling-walls-2023/https://laion.ai/blog/falling-walls-2023/We happily announce that we, LAION, won the Falling Walls Science Breakthrough of the Year 2023 Award in the category Science and Innovation Management for "democratizing AI research by providing open access to advanced AI models, tools, and datasets, fostering public engagement and awareness, and promoting internation…Thu, 14 Sep 2023 00:00:00 GMTChristoph, Jenia, Roberthttps://laion.ai/blog/visit_bench/https://laion.ai/blog/visit_bench/[[Paper]](https://arxiv.org/abs/2308.06595) [[Code]](https://github.com/mlfoundations/VisIT-Bench/) [[Dataset]](https://huggingface.co/datasets/mlfoundations/VisIT-Bench) [[Leaderboard]](https://huggingface.co/spaces/mlfoundations/VisIT-Bench-Leaderboard) We are thrilled to introduce VisIT-Bench , a benchmark for evalu…Tue, 15 Aug 2023 00:00:00 GMTYonatan Bittonhttps://laion.ai/blog/objaverse-xl/https://laion.ai/blog/objaverse-xl/We are thrilled to announce Objaverse-XL, an open dataset of over 10 million 3D objects! Using it, we train Zero123-XL, a foundation model for 3D that displays remarkable generalization abilities. In the landscape of AI, scale has been paramount to recent advances. Over the past decade, we have observed an escalating t…Tue, 11 Jul 2023 00:00:00 GMTMatt Deitkehttps://laion.ai/blog/video2dataset/https://laion.ai/blog/video2dataset/[[GitHub]](https://github.com/iejMac/video2dataset) Within only two years large foundational models like CLIP, Stable Diffusion, and Flamingo have fundamentally transformed multimodal deep learning. Because of such models and their impressive capabilities to either create stunning, high-resolution imagery or to solve c…Mon, 10 Jul 2023 00:00:00 GMTMaciej Kilianhttps://laion.ai/blog/open-flamingo-v2/https://laion.ai/blog/open-flamingo-v2/[[GitHub]](https://github.com/mlfoundations/open flamingo) [[Demo]](https://huggingface.co/spaces/openflamingo/OpenFlamingo) [[Models]](https://huggingface.co/openflamingo) About three months ago, we announced OpenFlamingo, an open-source effort to replicate DeepMind's Flamingo models. Today, we are excited to release…Wed, 28 Jun 2023 00:00:00 GMTAnas Awadalla* and Irena Gao*https://laion.ai/blog/datacomp/https://laion.ai/blog/datacomp/[[ Paper ]](https://arxiv.org/abs/2304.14108) [[ Code ]](https://github.com/mlfoundations/datacomp) [[ Website ]](http://datacomp.ai/) About a year ago, we released LAION-5B, a billion-scale open-source image-text dataset. Since then, LAION-5B has become a staple in the open-source machine learning ecosystem, powering…Thu, 27 Apr 2023 00:00:00 GMTGabriel Ilharcohttps://laion.ai/blog/paella/https://laion.ai/blog/paella/Overview. We are releasing a new Paella model which builds on top of our initial paper https://arxiv.org/abs/2211.07292. Paella is a text-to-image model that works in a quantized latent space and learns similarly to MUSE and Diffusion models. Paella is similar to MUSE as it also works on discrete tokens, but is differe…Sat, 15 Apr 2023 00:00:00 GMTDominic Rampas and Pablo Perniashttps://laion.ai/blog/petition/https://laion.ai/blog/petition/LINK TO OUR PETITION Authors: Christoph Schuhmann, Huu Nguyen, Robert Kaczmarczyk, Jenia Jitsev & LAION community Securing Our Digital Future: Calling for CERN like international organization to transparently coordinate and progress on large-scale AI research and its safety In an era of unparalleled technological advan…Wed, 29 Mar 2023 00:00:00 GMTLAION.aihttps://laion.ai/blog/open-flamingo/https://laion.ai/blog/open-flamingo/Overview. We are thrilled to announce the release of OpenFlamingo, an open-source reproduction of DeepMind's Flamingo model. At its core, OpenFlamingo is a framework that enables training and evaluation of large multimodal models (LMMs). Check out our GitHub repository and demo to get started! For this first release, o…Tue, 28 Mar 2023 00:00:00 GMTAnas Awadalla and Irena Gaohttps://laion.ai/blog/oig-dataset/https://laion.ai/blog/oig-dataset/The Open Instruction Generalist (OIG) dataset is a large open source instruction dataset that currently contains 43M instructions. OIG is one of many chatbot datasets that LAION, along with its volunteers, Ontocord, Together and other members of the open source community, will be releasing and is intended to create equ…Fri, 10 Mar 2023 00:00:00 GMTBy Huu Nguyen - Ontocord.ai, Sameer Suri, Ken Tsui , Shahules786, Together.xyz team, and Christoph Schuhmann - LAION.aihttps://laion.ai/blog/coca/https://laion.ai/blog/coca/We introduce a new model type to OpenClip Contrastive Captioners (CoCa) [1]. This model adds an autoregressive objective (generation) on top of the CLIP contrastive one. The architecture is composed of three parts, the first two are similar to those composing a CLIP model and the third is a text decoder that stands on…Thu, 02 Feb 2023 00:00:00 GMTGiovanni Puccetti, Maciej Kilian, Romain Beaumonthttps://laion.ai/blog/h14_clip_retrieval/https://laion.ai/blog/h14_clip_retrieval/Today we release a KNN index for LAION-5B that allows for fast queries of the dataset with the open clip ViT-H-14 CLIP model. This means that users can search through billions of samples quickly and easily, making it a powerful tool for various applications such as image and text retrieval, data filtering and more. Wit…Tue, 31 Jan 2023 00:00:00 GMTno usrhttps://laion.ai/blog/giant-openclip/https://laion.ai/blog/giant-openclip/We have trained a new ViT-G/14 CLIP model with OpenCLIP which achieves 80.1% zero-shot accuracy on ImageNet and 74.9% zero-shot image retrieval (Recall@5) on MS COCO. As of January 2023, this is the best open source CLIP model. We believe this is interesting because: CLIP models are useful for zero-shot classification,…Tue, 24 Jan 2023 00:00:00 GMTMitchell Wortsmanhttps://laion.ai/blog/laion-stable-horde/https://laion.ai/blog/laion-stable-horde/Author: Konstantinos Thoukydidis, hlky We are happy to announce that LAION will be assisted by the Stable Horde to provide aesthetic ratings for existing datasets and a completely new dataset of Stable Diffusion generations, which will also be rated by their community. We wrote in the past about LAION-Aesthetics and ho…Sun, 08 Jan 2023 00:00:00 GMTKonstantinos Thoukydidis, hlkyhttps://laion.ai/blog/laion-coco/https://laion.ai/blog/laion-coco/Author: Christoph Schuhmann, Andreas Köpf , Theo Coombes, Richard Vencu, Benjamin Trom , Romain Beaumont We present LAION-COCO, the world’s largest dataset of 600M generated high-quality captions for publicly available web-images Laion5B has five billion natural captions. They provide a lot of information, but could sy…Thu, 15 Sep 2022 00:00:00 GMTChristoph Schuhmann, Andreas Köpf, Richard Vencu, Theo Coombes, Romain Beaumonthttps://laion.ai/blog/laion-translated/https://laion.ai/blog/laion-translated/Author: Marianna Nezhurina Romain Beaumont Richard Vencu Christoph Schuhmann Laion5B dataset was automatically collected from a section of the human web (common crawl). Can models generate different and interesting data compared to what humans write? That’s a question we are interested in investigating. To let the comm…Thu, 15 Sep 2022 00:00:00 GMTMarianna Nezhurina, Romain Beaumont, Richard Vencu and Christoph Schuhmannhttps://laion.ai/blog/large-openclip/https://laion.ai/blog/large-openclip/We trained three large CLIP models with OpenCLIP: ViT-L/14, ViT-H/14 and ViT-g/14 (ViT-g/14 was trained only for about a third the epochs compared to the rest). The H/14 model achieves 78.0% zero shot top-1 accuracy on ImageNet and 73.4% on zero-shot image retrieval at Recall@5 on MS COCO. As of September 2022, this is…Thu, 15 Sep 2022 00:00:00 GMTRomain Beaumonthttps://laion.ai/blog/laion-aesthetics/https://laion.ai/blog/laion-aesthetics/We present LAION-Aesthetics, several collections of subsets from LAION 5B with high visual quality. To create LAION-Aesthetics we trained several lightweight models that predict the rating people gave when they were asked “How much do you like this image on a scale from 1 to 10?” . LAION-Aesthetics V1 We started with t…Tue, 16 Aug 2022 00:00:00 GMTChristoph Schuhmannhttps://laion.ai/blog/laion-5b/https://laion.ai/blog/laion-5b/We present a dataset of 5,85 billion CLIP-filtered image-text pairs, 14x bigger than LAION-400M, previously the biggest openly accessible image-text dataset in the world - see also our NeurIPS2022 paper See our update on the LAION-5B dataset. Large image-text models like ALIGN, BASIC, Turing Bletchly, FLORENCE & GLIDE…Thu, 31 Mar 2022 00:00:00 GMTRomain Beaumonthttps://laion.ai/blog/laion-400-open-dataset/https://laion.ai/blog/laion-400-open-dataset/We present LAION-400M: 400M English (image, text) pairs - see also our Data Centric AI NeurIPS Workshop 2021 paper Concept and Content The LAION-400M dataset is entirely openly, freely accessible. WARNING : be aware that this large-scale dataset is non-curated. It was built for research purposes to enable testing model…Fri, 20 Aug 2021 00:00:00 GMTChristoph Schuhmann