Welcome to AI Current. Today: how OpenAI is stress-testing models before they ship, a new open-weight coding model out of China, Google's latest Pixel update, and a cluster of business moves reshaping how AI is priced and sold. Here's what happened over the last 24–48 hours →

A word from our sponsor

Top engineers at Anthropic and OpenAI say AI now writes 100% of their code.

If you're not using AI, you're spending 40 hours doing what they do in 4.

These 100+ Claude Code hacks fix that and help you ship 10x faster.

Sign up for The Code and get:

Claim your free playbook

OpenAI details “Deployment Simulation” for pre-release testing. OpenAI described a method that replays large volumes of anonymized past conversations through a candidate model before release to estimate how it will behave once live. The company says it pre-registered predictions for 20 categories of undesirable behavior and analyzed roughly 1.3 million de-identified conversations spanning models from GPT-5 Thinking through GPT-5.4, reporting a median multiplicative error of about 1.5x against later measurements. OpenAI frames the approach as a complement to adversarial red-teaming rather than a replacement. Read more →

Zhipu AI releases GLM-5.2, an open-weight coding model. The Chinese lab released GLM-5.2, a coding-focused model with a 1-million-token context window distributed under an MIT license, with open weights slated to publish on Hugging Face the week of June 16. Zhipu is positioning the model for agentic coding workflows. It launched without an accompanying benchmark report, so independent comparisons remain pending. Read more →

Google ships the June 2026 Pixel Drop alongside Android 17. Google began rolling out its June Pixel Drop with the stable release of Android 17. The update adds Bubbles, a multitasking feature that turns apps into floating overlays, plus several AI tools: Gemini Omni 2 for text- and photo-to-video creation, music generation through Gemini, and Screen Reactions for recordings. Magic Cue, still limited to the Pixel 10, is set to gain support for additional messaging apps in the coming weeks. Read more →

xAI adds an Agent Dashboard to Grok Build. xAI introduced an Agent Dashboard that gives developers a single screen to start and manage multiple coding sessions at once. The feature ships as part of the Grok Build developer stack. Read more →

A word from our sponsor

The best HR advice comes from those in the trenches. That’s what this is: real-world HR insights delivered in a newsletter from Hebba Youssef, a Chief People Officer who’s been there. Practical, real strategies with a dash of humor. Because HR shouldn’t be thankless—and you shouldn’t be alone in it.

Sign Up Free

OpenAI launches a $150M Partner Network. OpenAI announced its first formal partner program, backed by a $150 million investment and a goal of certifying 300,000 consultants by the end of 2026. The program has three tiers—Select, Advanced, and Elite—with specializations in areas such as Codex, cybersecurity, and agent deployment, and is scheduled to go live in July. Launch participants include Accenture, Bain, BCG, McKinsey, and PwC. OpenAI framed the effort around enterprise adoption hurdles rather than raw model capability. Read more →

Anthropic pauses a planned Agent SDK billing change. Anthropic had told subscribers that, starting June 15, programmatic Claude usage—including the Agent SDK, the headless claude -p command, and third-party apps—would draw from a separate monthly credit pool billed at API rates rather than from Claude subscription limits. On June 15 the company confirmed it was pausing the change, so subscription-based access for those use cases continues for now. The episode highlighted how usage-based pricing interacts with flat-rate plans for agent workloads. Read more →

Anthropic works through repeated Claude outages. Anthropic logged another Claude service disruption on June 16 that trade coverage described as the tenth in roughly 12 days, with errors affecting Opus 4.8 and Haiku 4.5 persisting after an initial fix. The company addressed the incidents through rolling status-page updates rather than a standalone statement. The cluster of outages has drawn attention to infrastructure capacity amid rising demand. Read more →

That's today's AI Current. We aim to keep this factual and concise—reply with a correction if we got something wrong, or forward it to a colleague who'd find it useful.

— The AI Current team