Open-Source AI
Open models — Llama, Mistral, DeepSeek, Qwen and others — are catching up to closed labs and let you run AI on your own hardware, vendor-free. That shifts the balance of power, the safety debate and the economics of the field. Here is everything on open-source AI: open-weight releases, licenses, local deployment and the race against closed models.
Google DeepMind released DiffusionGemma — an open 26B MoE model with 4x faster generation
Google DeepMind released DiffusionGemma, an experimental 26B-parameter open model that uses text diffusion instead of autoregression — generation is up to 4x faster…
PyGraphistry: Graph analytics for threat detection in enterprise data
Researchers published a PyGraphistry Colab workflow that turns enterprise access data into an interactive graph with ML-based anomalies and three visualization…
Google DeepMind releases DiffusionGemma — a diffusion-based LLM 4x faster than other Gemma 4 models
Google DeepMind unveiled DiffusionGemma, the first diffusion-based language model in the Gemma lineup, running 4x faster than other Gemma 4 models at comparable…
JetBrains released Mellum2 — an open 12B MoE model for fast tasks in AI pipelines
NVIDIA JetPack 7.2: AI agents on edge devices with a single deployment command
Meta explains the difference between GPU, CPU, and its own MTIA chips in AI infrastructure
LangChain released Promptim, a library for automatic prompt optimization
Google DeepMind's Gemma 4 arrives on Amazon Bedrock: three models with MoE and multimodality
Amazon Bedrock has added three open-weight Gemma 4 models from Google DeepMind, featuring MoE architecture, built-in reasoning, and multimodal input support.
Mistral released OCR 4: bounding boxes, 170 languages, and self-hosted deployment
Mistral AI released OCR 4 — an engine for enterprise document processing with block coordinates, type classification, and support for 170 languages in a single…
DeepSeek released DSpark — a framework that speeds up generation on DeepSeek-V4 by 57–85%
Meta Unveils Astryx — A React Design System with AI Agent and MCP Support
Liquid AI released LFM2.5-230M: 213 tokens/s on Galaxy S25 and support for llama.cpp
graphlens-mcp builds a code graph with one command and cuts token usage by 10–23x
GitHub Joins Coalition Against AI Transparency Law's Conflict With Open Source
GitHub has joined a coalition pushing for amendments to California's AI transparency law: as currently written, it conflicts with open-source licenses.
ML Red Teaming for LLMs: From Hallucinations to Data Leaks — Testing in Practice
How to attack LLM models to find vulnerabilities before adversaries: a practical breakdown of attack classes, testing methodologies, and defenses for enterprise AI.