Claude Opus 5: Frontier Performance at Half the Price of Fable 5
Anthropic launched Claude Opus 5 on July 24, 2026: near-Fable-5 benchmark scores at half the cost. Here is what changed, how the new ...
Vercel AI SDK 7: WorkflowAgent for Production Agents
Vercel AI SDK 7 ships WorkflowAgent for durable agent execution, typed tool context, layered timeouts, ...
Gemini API Managed Agents: Background Execution and Remote MCP Guide
Google's July 7 Gemini API update adds background execution, remote MCP, and credential refresh to ...
Claude Code Subagent Guardrails: Three Limits, Four Days
Anthropic shipped hard subagent limits across four days in July: concurrency 20, session cap 200, ...
IntelliJ IDEA 2026.2: GitHub Copilot Goes Native
IntelliJ IDEA 2026.2 makes GitHub Copilot a native agent — no plugin setup, no ACP ...
NVIDIA Nemotron-Labs TwoTower: 2.42x Faster Inference, No Retraining Required
NVIDIA Nemotron-Labs TwoTower retrofits a pretrained 30B AR model into a diffusion decoder running 2.42x faster at 98.7% quality. Open weights, commercial license, ...
Buzz by Block: Open-Source Workspace Where AI Agents Own Their Actions
Block launched Buzz on July 21 — an Apache 2.0 workspace where AI agents join ...
Copilot Cloud Agent for Linear Is Now GA: Setup and Limits
GitHub's Copilot cloud agent for Linear is GA. Assign an issue, get a PR. Here ...
AI Is Buying Rare Books and Shredding Them. Here’s Why
AI companies are bulk-buying rare books, scanning them, and shredding the originals. The reason: their ...
Kimi K3 Open Weights Live: Self-Host or Use the API?
Kimi K3’s 594 GB MXFP4 weights are live on Hugging Face under Apache 2.0. Here’s ...