News

For about two weeks in June, the most capable AI model on the market was, in a real legal sense, off-limits. If you were not a US national, you were not allowed to touch it. This is the Claude Fable 5 story, and it is less about one model than about a line that just got crossed.

At WWDC in June, Apple did the thing everyone had demanded for two years: it tore Siri down to the studs and rebuilt it. The new Siri AI is genuinely ambitious. It is also, on closer reading, a promise with a lot of asterisks.

DeepSeek V4 arrived in April with the same trick DeepSeek always pulls: numbers that should cost a fortune, at prices that do not. Top-tier coding scores, a million-token context, an MIT license, and output priced under a dollar per million tokens. The West still has not quite made peace with it.

Every month brings a new open model, and most are footnotes. GLM-5.2 is not. Z.ai released a 753-billion-parameter model, with a one-million-token context window, under a plain MIT license. That last part is what makes it matter.

Meta Llama 4 family landed with a headline number that is hard to ignore: Scout, the smaller variant, offers a ten-million-token context window. That is the largest of any model, open or closed. It is also a feature most people quoting it will never actually use.

For years the deal was simple: Microsoft paid, OpenAI built, and Copilot ran on someone else models. At Build 2026, Microsoft quietly changed the arrangement by shipping seven models of its own. The most important AI news of the month was not a benchmark. It was a message.

Moonshot released Kimi K2.7 Code, a trillion-parameter coding model, and buried in the announcement was the number that actually matters: it reaches its answers using roughly 30 percent fewer reasoning tokens than its predecessor. In a world obsessed with capability scores, efficiency is the upgrade people underrate, and it is quietly the more useful one.

Most model releases are the same recipe made a little bigger. NVIDIA Nemotron 3 Ultra is a bit more interesting, because it quietly does something different under the hood. It is a 550-billion-parameter model built on a hybrid of Mamba and Transformer architectures, and that architectural choice is the actual story, not the benchmark line.

OpenAI began rolling out GPT-5.6 in early July, loosening some of the access limits that had been in place. The upgrades are genuine. The name is a small crime against anyone trying to keep a mental map of which model does what.

Quietly, without a single headline capturing it, the cost of using capable AI collapsed this year. Not dropped. Collapsed. The kind of model that cost a small fortune to run at scale two years ago now costs cents per million tokens. If your business plan assumed AI would stay expensive, it is time to redo the math.

June and early July gave us a model release almost every day, which is exactly why you should not try to follow all of them. Most were incremental. A few will still matter next summer. Here is the month sorted the only way that is useful: by how long it will stay relevant.

xAI reportedly acquired the company behind Cursor, the AI coding editor, for $60 billion in all stock. You can argue about the number all day. What you cannot argue with is the message it sends: the smart money has decided the value in AI is no longer in the model. It is in the thing wrapped around it.