Where the AI industry shifts next.

4 Checks Before Deploying GPT-6 Astra, the First 'Critical' Cyber Model

Anthropic Releases Claude Opus 5 as Its New Default for Max

Anthropic introduced Claude Opus 5, describing stronger coding and knowledge-work performance at the same price as its predecessor.

Meta Launches Muse Image and Previews Muse Video

Meta introduced its first Superintelligence Labs media models, pairing an available image model with an early video preview.

AI Agent Architecture Explained

What Is an AI Agent?

Mistral Launches Agentic Search for Complex Enterprise Documents

Mistral introduced a retrieval layer that lets models search, inspect and verify information across complex data sources.

Mistral Expands Regional Inference and European AI Compute Plans

Mistral outlined regional endpoints, third-party open-model support and a roadmap for up to one gigawatt of European AI capacity.

Anthropic Rejects a Blanket Ban on Open-Weights Models

Anthropic said it did not support a categorical ban, while arguing that sufficiently capable open models should face safety testing.

Mistral Releases OCR 4 for Document Intelligence and Retrieval Pipelines

Mistral launched OCR 4 with layout data, typed blocks and confidence scores for search, RAG and document-processing workflows.

$900M ARR and a $48B Bet: What Cognition's Round Says About Agent Demand

Microsoft Puts Day-Long Agent Work Inside a Windows 365 Cloud PC

15,000 Agent Edits, One EU Filing: OpenAI's Wiki Incident Meets Article 55

EU AI Act Enforcement Starts: 30+ AI Firms Now Face Formal RFIs

Audit .git/config Now: GitSpawn Runs Attacker Code in 4 Unpatched AI Agents

Mistral Releases Shieldstral, a Small Open-Weights Safety Classifier

Mistral released Shieldstral, a 3B multimodal classifier that applies plain-language policies at inference time.

OpenAI Launches Prism, an AI-Native Workspace for Scientific Writing

OpenAI introduced Prism as a free collaborative research-writing workspace powered by GPT-5.2.