So excited to be the SFT lead of this massive collaboration! The OpenThoughts team is 🔥

If you are attending #ICML2025, check out our DataWorld workshop on Sat July 19. We have updated the website with more info on speakers & accepted papers! dataworldicml2025.github.io/ Also happy to chat offline about all things ✨ data ✨

New research paper for you to read over your July 4th break (if you're US-based) -- Vision is a skeleton key! 🗝️ We convert a small VLM into an "everything classifier" by transforming data into visualizations that VLMs can naturally understand and reason about. We call it MARVIS: Modality Adaptive Reasoning over VISualizations. Our MARVIS-3B model: - Beats Gemini by 16% on average across 100s of vision and tabular tasks 🏆 - Gets within 2.5% of the best specialized model across across 4 modalities ... 🎯 - Using just one 3B model ... 💪 - ... without exposing any P.I.I. (personally identifiable information) to the VLM ... 🔐 - And without requiring any model training! ⚡ Our GitHub: lnkd.in/eqPkzU2m 💻 Our Paper: lnkd.in/esXZEvEE 📄 Research Supported By: oumi.ai Thanks to @LennartPurucker @Oussama_e

Announcing OpenThinker3-7B, the new SOTA open-data 7B reasoning model: improving over DeepSeek-R1-Distill-Qwen-7B by 33% on average over code, science, and math evals. We also release our dataset, OpenThoughts3-1.2M, which is the best open reasoning dataset across all data scales. Full details are in our ✨new paper✨ - below we share the highlights: BTW, it also works on non-Qwen models😉 (1/N)

📣We are extending our deadline to May 31st!📣 Looking forward to seeing everyone's submissions :)

Many agents (Claude Code, Codex CLI) interact with the terminal to do valuable tasks, but do they currently work well enough to deploy en masse? We’re excited to introduce Terminal-Bench: An evaluation environment and benchmark for AI agents on real-world terminal tasks. Tl;dr lots of room for improvement! tbench.ai/

📢 Announcing our data-centric workshop at ICML 2025 on unifying data curation frameworks across domains! 📅 Deadline: May 24, AoE 🔗 Website: dataworldicml2025.github.io/ We have an amazing lineup of speakers panelists from various institutions and application areas.

If AI isn’t truly open, it will fail us. We can’t close in a black box our greatest invention yet just so that a few can freely monetize. AI needs its Linux moment, and so we started working towards it. This can only succeed if we all work together! #oumi #opensource #collaboration @rsalakhu @svlevine @larry_heck @karpathy @atalwalkar @prfsanjeevarora @tsvetshop @hhexiy @sainingxie @larry_heck @AnimaAnandkumar @JunjieHu12 @georgiagkioxari @profjoeyg @pliang279 @danqi_chen @ChrisGPotts @BillMacCartney @vinodv @tur_gokhan @dilekhakkanitur @j_foerst @gingsmith23 @kahinish @jamesjoaquin @gan3sh @ethanjb @kirbywinfield @egonzdp

Oumi:build state-of-the-art foundation models, end-to-end. Oumi is a fully open-source platform designed to train, evaluate, and deploy foundation models end-to-end. It supports models from 10M to 405B parameters, enabling fine-tuning using LoRA, QLoRA, DPO, and other techniques. It integrates with popular inference engines (vLLM, SGLang) and works across laptops, clusters, and cloud platforms (AWS, Azure, GCP, Lambda, etc.). Supports multimodal models like Llama, DeepSeek, and Phi. ⚙️ Key Benefits → Oumi simplifies model training with a unified API, allowing seamless model fine-tuning, data synthesis, and evaluation. Supports both open-source and commercial APIs like OpenAI, Anthropic, and Vertex AI, making it highly flexible. → Enables fast inference with optimized engines such as vLLM and SGLang, ensuring efficient deployment. Installation is straightforward with pip install oumi, supporting both CPU and GPU setups. → Features a CLI tool (oumi train, oumi evaluate, oumi infer) for easy model training, evaluation, and inference. → Supports cloud-based training with direct job execution on AWS, Azure, GCP, and Lambda. → Includes prebuilt ready-to-use training recipes for LLM fine-tuning, distillation, evaluation, and inference. → 100% open-source under Apache 2.0 license, with an active community on Discord and GitHub.

Oumi is a fully open-source platform to help you build state-of-the-art foundation models, end-to-end.

After careful consideration, I have decided to leave X for BlueSky. I hope to see many of you there with me very soon! @benjaminfeuer.bsky.social

🐞 Check out this zero-shot AI model dataset with 6M images of important species that is vital to farming and environmental research! Learn more: ow.ly/Kzig50TRSaC #ZeroShotLearning #AIModels #AgriculturalTech #EnvironmentalResearch @chegday @FeuerBenjamin @AII4RA