It's been a dense few days in AI, even by 2026 standards. Three storylines stand out: OpenAI reshaping how ChatGPT looks and behaves, a brewing credibility fight between OpenAI and the mathematics establishment, and Anthropic firing another shot in the ongoing AI pricing war.

ChatGPT stops being just text

On October 7, OpenAI began rolling out GPT-6 to the entirety of ChatGPT's user base alongside a new feature called Intelligent UI. Instead of returning plain paragraphs, GPT‑6 in ChatGPT introduces Intelligent UI, a capability that allows ChatGPT to answer with fully interactive user interfaces that make everyday answers more visual and can quickly create a tool to solve a task right in the moment. The rollout targets a staggering audience: OpenAI said the new GPT-6 model in ChatGPT is built for the more than 1.2 billion people who use ChatGPT each week.

The practical effect is that a question about retirement savings might now surface as a live calculator rather than a paragraph of advice, and splitting a dinner bill could spawn an actual interactive tool inside the chat window. Paid tiers (Plus, Pro, Business, Enterprise) got the feature first, powered by the mid-tier "Sol" model, while free and Go users received the lighter "Luna" model a day later. Notably, the update applies to the Chat experience in ChatGPT, and the models powering Work and Codex are not changing as part of this release. It's a clear bet that the next competitive battleground in consumer AI isn't just model intelligence but interface — turning a chatbot into something closer to a dynamic app generator.

OpenAI's math blitz reopens an old wound

Just a day earlier, OpenAI dropped a very different kind of release: a massive GitHub dump of AI-generated mathematics. On October 6, OpenAI released 722 manuscripts organized into 372 research families, all credited to an internal model the company has not named, built on the Lean proof assistant so every result comes with a machine-checkable certificate, posed roughly 4,000 open problems, with each surviving result taking about three hours of ChatGPT Pro compute on average.

The timing matters. Three days after OpenAI's earlier Navier-Stokes announcement, 25 Fields Medalists signed a declaration titled "A Severe Misalignment of AI in Mathematics," not disputing that the proofs checked out but objecting to results landing as press releases instead of papers, with no named authors or attribution trail. Compounding the unease, Fields Medalist Terence Tao wrote that problems are being solved by "AI prompters" who do not understand the output well enough to answer questions or give talks on it. It's a striking tension: genuinely useful machine-generated proofs colliding with a scholarly culture built on peer review and attribution.

The small-model price war keeps escalating

Meanwhile, Anthropic quietly undercut everyone on cost. On October 7 it launched Claude Haiku 5.5, and cutting token prices by 90% for requests below 100,000 tokens while targeting repetitive enterprise work like summarizing documents, classifying information and querying databases, with pricing starting at $0.10 per million input tokens and $0.50 per million output tokens — matching OpenAI's GPT-6 Luna. Anthropic also added adjustable effort controls and a sharply larger context window, with Haiku 5.5 now carrying a one-million-token context window, up from the 200,000 tokens Haiku 4.5 shipped with a year ago.

Taken together, these three stories sketch where the industry is headed: richer, more interactive products at the consumer layer; mounting friction over how AI-generated research gets validated and credited; and a relentless economic squeeze that's making frontier-adjacent AI dramatically cheaper to deploy at scale.

Sources

AI ResearchAnthropicArtificial IntelligenceChatGPTDaily AI UpdateOpenAI