Models, access, ideas to build on. A place to discover developments quickly, then stay with the questions that matter to you.
Newest source publications first. These are source dates, not live activity.
Foundation-Sec-8B is a cybersecurity-focused Llama 3.1 derivative that organizations can download and adapt for security operations work involving their own alerts, cases, and threat knowledge.
Emerging · Cisco Foundation AINVIDIA announced the Llama Nemotron Nano and Super reasoning models, alongside plans to make the post-training tools, datasets and optimization techniques openly available.
Emerging · NVIDIAAi2 released OLMo 2 32B, a language model accompanied by publicly available data, code, weights, training details, and a reproducible training recipe.
Demonstrated · Ai2NVIDIA and partners launched Signs, an interactive platform for ASL learning and a validated sign-language dataset.
Emerging · NVIDIA / American Society for Deaf Children / Hello MondayMatterGen generates candidate inorganic crystal structures while conditioning on properties a researcher wants, changing the starting point from searching a known catalog to proposing structures for testing.
Demonstrated · Microsoft ResearchPrithvi-EO-2.0 is a geospatial foundation model that can be adapted to Earth-observation tasks such as flood mapping, crop classification, and land-use analysis.
Demonstrated · NASA and IBM ResearchSmolVLM is a compact vision-language model family designed for document, image, and visual-question tasks where memory and compute are limited.
Emerging · Hugging FaceBioNeMo brings biological foundation models and training components into one framework for protein, molecular, and drug-research experiments.
Emerging · NVIDIAOpenVLA gives robotics teams a pretrained vision-language-action model plus instructions for fine-tuning it on their own robot demonstrations.
Experimental · Stanford, UC Berkeley, and Toyota Research Institute collaboratorsOcto is a generalist manipulation policy trained on the Open X-Embodiment dataset and evaluated as a reusable starting point for robots with different sensors, action spaces, and physical forms.
Demonstrated · UC Berkeley, Stanford University, Carnegie Mellon University, and Google DeepMind collaboratorsStarCoder2 offers several downloadable code-model sizes plus fine-tuning and evaluation examples for local completion and generation experiments.
Emerging · BigCodeMedSAM adapts promptable segmentation to medical images and provides interactive tools for delineating structures or lesions for research annotation workflows.
Experimental · University research collaborators led by the University of Toronto