Google DeepMind Releases EmbeddingGemma 2, a 740M Open Multimodal Embedding Model Built on Gemma 4
Source: MarkTechPost Google DeepMind has released EmbeddingGemma 2, an open model that embeds text, code, images, video and...
Mistral AI Releases Mistral Large 4 (Le Chonk): A 1.05T Parameter Multimodal MoE Model
Source: MarkTechPost Mistral AI has just announced the release of Mistral Large 4 (ML4), internally nicknamed Le Chonk,...
Reka Releases Rho-1: A 19B Omni-Reasoning Model That Understands, Generates Video and Outputs Robot Actions in One
Source: MarkTechPost Reka has released a research preview of Rho-1, a 19B omni-reasoning model trained from scratch. A...
Beyond Domain-Specific World Models: JEPA-Anything Uses 1 Recipe for 7 Fields
Source: MarkTechPost Researchers from PhAI Labs, CUHK, Fudan, Stanford, Oxford and Princeton have released JEPA-Anything, a domain-agnostic framework...
Meet Together Link: A Free CLI That Runs Open Models Like Kimi K3 and GLM 5.3 Inside Claude Code, Codex, and OpenCode
Source: MarkTechPost Together AI has released Together Link, a free, MIT-licensed CLI now in beta. It connects the...
Building a Streaming Robotics Learning Pipeline Using NVIDIA Cosmos3-DROID
Source: MarkTechPost In this tutorial, we design an end-to-end streaming robotics learning pipeline around the NVIDIA Cosmos3-DROID dataset...
Reflection AI Introduces Beam: A 501B Open-Weight MoE Model With 23B Active Parameters for Coding and Agentic Workloads
Source: MarkTechPost Reflection AI has introduced Beam, its first open-weight model. Beam is a sparse Mixture-of-Experts (MoE) model...
Yandex Introduces Sona: A Single Generative Recommender That Replaces Entire Recommendation Cascade
Source: MarkTechPost Most production recommenders are cascades. Candidate generators feed a pre-ranker, which feeds a heavy ranker built...
The Story of Qwen: Alibaba’s AI Models From 7B to 2.4T
Source: MarkTechPost In April 2023, Alibaba Cloud demoed a chatbot whose name roughly means ‘truth from a thousand...
Can an Open Model Do Security Research? Cantina’s apex-flash-1 Solves 40 of 60 Held-Out Bug Tasks
Source: MarkTechPost Cantina Security, with Yeta Labs, has released apex-flash-1, an open-weights model trained specifically for vulnerability research....