OpenAI Holds Back GPT-5.6 for a Government-Approved Few
OpenAI limited its new GPT-5.6 models to a small group of partners the government signed off on โ the second top lab in two weeks to gate its best model behind a federal review.
OpenAI limited its new GPT-5.6 models to a small group of partners the government signed off on โ the second top lab in two weeks to gate its best model behind a federal review.
The government made Anthropic pull Fable 5 and Mythos 5 for everyone, over a jailbreak that amounts to asking a model to fix bugs in code. The facts, and why it hit me personally.
Two days into the Fable 5 window, a binary export of blank mammography templates tripped Anthropic's new safety classifiers โ cybersecurity and biology at once. Then writing this post tripped them again. Both times, the quiet handoff to Opus 4.8 worked exactly as advertised.
Anthropic's most powerful public model matches or beats Opus across the board โ then quietly hands its riskiest queries back to Opus to keep them safe.
OpenAI has confidentially filed for a U.S. IPO, following Anthropic and SpaceX toward the public markets. With a debut possibly as early as September at up to a $1 trillion valuation, the real test is whether the market can absorb three of the largest offerings ever in a matter of months.
Claude blocked a routine web search about an npm vulnerability and pointed me to Anthropic's Cyber Verification Program. I applied out of curiosity โ and the approval came back so fast that a human almost certainly never read it.
Anthropic's new research says AI is already accelerating its own development, and that a model training its own successor could arrive before institutions are ready. Its proposed answer isn't a pause โ it's the machinery to make a coordinated slowdown verifiable.
Anthropic confidentially filed a draft S-1 with the SEC on June 1, taking the first formal step toward an IPO โ and beating OpenAI to it. What a confidential filing does and doesn't mean, and the question it leaves open: what happens to a safety-first public benefit corporation when public shareholders arrive.
Trump signed an order asking AI companies to voluntarily hand the government a 30-day look at their most powerful models before release. It explicitly bars any mandatory licensing โ making the whole thing a request, not a rule.
Anthropic closed its Series H at $65 billion, valuing the company at $965 billion โ past OpenAI's $852 billion mark and within reach of a trillion. The round that was still being weighed a month ago came in bigger than reported.