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.
A Year Later: Qwen3 Still Holds the Price/Quality Crown — LLM Model Battletest
Qwen3-235B from July 2025 once again tops the price/quality ratio. A year on, competitors like Gemini and DeepSeek failed to surpass it.
Microsoft Restricted Claude Fable 5 for Employees Due to Data Policy
Anthropic released a new Claude Fable model, but Microsoft banned its internal use due to data storage requirements.
Google introduces DiffusionGemma — a model with an alternative approach to text generation
Google released an experimental DiffusionGemma model that uses a diffusion approach instead of a transformer — a fundamentally different text generation mechanism…
NVIDIA X-Token: distillation that beats GOLD by 3.82 points
AWS SageMaker adds full monitoring for LLM models in production
Groq raises $650 million and shifts from chips to an inference platform
MiniMax Prepares for IPO in China, Competing With Fast-Growing DeepSeek
Ollama Instead of ChatGPT: Saving Money, Privacy, and the Planet
A developer ditched ChatGPT for Ollama — a free local AI that works without internet, subscriptions, or data tracking.
Mistral Medium 3.5: Cloud Agents and New Work Mode for Complex Tasks
Mistral AI launched a new 128B Medium 3.5 model and moved coding agents to the cloud—now they work asynchronously and in parallel while developers focus on other…
Mistral Launches Search Toolkit for AI Applications
NVIDIA's New CompileIQ Unlocks Hidden GPU Core Potential Through Compiler Parameter Tuning
Bank of Italy Openly Announced Negotiations with All Major AI Companies
Anthropic Releases Claude Opus 4.8 with Improvements for Code and Agents
Mistral Released Voxtral TTS — Lightweight Text-to-Speech Model for Voice Agents
Mistral AI unveiled Voxtral TTS — a compact text-to-speech model that generates emotional speech in 9 languages and adapts to any voice in 3 seconds.
Mistral AI Renamed Le Chat to Vibe — a Universal AI Agent for Work and Code
Mistral unveiled Vibe — a unified AI agent that performs multi-step work tasks, manages context, and writes code with GitHub and VS Code integration.