This Week in AI: GPT-5.5, Grok 4, and $60B Valuations
A quick look at recent AI releases including OpenAI's GPT-5.5, Anthropic's quality transparency, and xAI's massive Cursor acquisition rumors.
Loading content...
A quick look at recent AI releases including OpenAI's GPT-5.5, Anthropic's quality transparency, and xAI's massive Cursor acquisition rumors.
This week brought major model updates with OpenAI releasing the GPT-5.5 series and xAI launching Grok 4. We also saw Anthropic admit to quality issues and reports emerge that xAI is looking to acquire Cursor at a staggering valuation.
A quick look at a few things that happened in the AI space over the last week.
OpenAI officially released the GPT-5.5 series this week. The rollout includes the standard GPT-5.5 model along with new Thinking and Pro variants. These models focus on agentic coding and computer-use accuracy, which you can read about in a recent coding benchmark comparing it with Opus 4.7. They also ship with a massive one million token context window.
There are reports that xAI is looking to acquire Cursor. Cursor was recently in talks for a funding round that would value the company at over $50 billion. Rumors indicate that SpaceX and xAI have secured an option to acquire the AI coding startup for $60 billion later this year. If the acquisition falls through, they would still pay $10 billion for a technology partnership. The numbers show how much value is being placed on controlling the developer experience.
In a rare move of transparency, Anthropic published an investigation regarding claims that Claude had "dumbed down". They admitted to three specific missteps. First, they lowered default reasoning effort to reduce latency. Second, a bug in cache optimization made the model forgetful. Third, a system prompt change made Claude too brief and less helpful.
They also reversed a controversial 2% test that had removed Claude Code from the $20 per month Pro plan.
Alongside the transparency report, Anthropic launched Claude Opus 4.7. This update improves performance for complex software engineering tasks. It also features a new "xhigh" effort level for deeper reasoning. On the product side, they introduced Claude Design to let users collaborate on creating visual outputs like prototypes and slides.
Meta announced it is laying off 10% of its workforce, which is roughly 8,000 employees. CEO Mark Zuckerberg cited the need to rebalance the company to support the massive infrastructure costs and personnel requirements of the post-GenAI era.
Elon Musk's xAI dominated the headlines by launching Grok 4 and its high-octane counterpart, Grok 4 Heavy. Musk claimed the model now rivals the intelligence of graduate students across all disciplines.
The Heavy tier is priced at $300 per month for power users. It reportedly utilizes five coordinated replicas to solve the most complex reasoning tasks.
It feels like the "honeymoon phase" of AI is shifting into a "performance and infrastructure" phase, where labs are being held accountable for model drift and companies are gutting traditional roles to pay for GPU clusters.