GLM-5.2: Open-Source Model Beats GPT-5.5 on SWE-bench for 1/6 the Cost
Z.ai’s GLM-5.2 hits 62.1 on SWE-bench Pro, topping GPT-5.5’s 58.6, with MIT weights on Hugging Face and API pricing at $1.40/M input tokens.
GitHub Agentic Workflows: No More PATs, Four Agent Engines
GitHub Agentic Workflows hit public preview on June 11. The PAT requirement is gone. Write ...
GitHub Copilot Metrics API: Track Per-User AI Spend Now
GitHub's Copilot usage metrics API now shows per-user AI credit consumption. Here's how to query ...
OpenAI Acquires Ona: Codex Gets an Enterprise Cloud
OpenAI acquired Ona (Gitpod) on June 11 to give Codex agents persistent cloud execution in ...
React Router v8: ESM-Only, Middleware On, Node 22 Now
React Router v8 landed on June 17. It requires Node 22, ships ESM only, and ...
WASI 0.3.0 Is Out: Native Async Kills the Polling Boilerplate
WASI 0.3.0 landed June 11 with native async at the ABI level. Here is what changed, what you need to migrate, and where ...
GitHub Actions Now Blocks Untrusted Workflow Triggers
GitHub shipped workflow execution protections and a safer checkout default on June 18. Here's what ...
Cursor v3.8: /automate Skill, Slack Triggers, and Computer Use
Cursor v3.8 shipped June 18 with changes that push it out of the IDE-copilot category ...
TanStack AI Beta: Code Mode, Middleware, and MCP Are Here
TanStack AI hits Beta with Code Mode sandbox execution, middleware, and host-side MCP. Here's what ...
Rust 1.95: cfg_select! Makes cfg-if Optional at Last
Rust 1.95 ships cfg_select!, a native stdlib macro that replaces the cfg-if crate for cross-platform ...