The AI platform wars heated up this week with two contrasting moves: Anthropic is expanding Claude's capabilities with specialized "domain expert" plugins, while OpenAI is cleaning house by retiring GPT-4o and several older models. Meanwhile, the biggest AI deal of 2025 appears to be unraveling.
Anthropic's Cowork Gets Specialized
Anthropic dropped a significant update to its Cowork tool — the collaborative workspace feature that turns Claude into more than just a chatbot. The new "plugins" system allows Cowork to function as a domain expert across specific fields: sales, legal, finance, marketing, data analysis, customer support, product management, and even biology research.
The concept is straightforward but powerful: instead of Claude being a generalist that's decent at everything, these plugins let it operate with the depth and context of a specialist. Think of it as the difference between asking a smart friend about tax law versus asking a tax attorney.
The feature is available now in research preview for all paid subscription tiers — a move that signals Anthropic is competing not just on model quality but on practical, work-ready tooling.
GPT-4o: End of an Era
In a move that says a lot about the pace of AI development, OpenAI announced it will retire GPT-4o, GPT-4.1, GPT-4.1 mini, and o4-mini from ChatGPT on February 13th.
The reason? Only 0.1% of users are still choosing GPT-4o on any given day. After GPT-5's launch, OpenAI had actually brought GPT-4o back because users missed it. But the nostalgia faded quickly — turns out, when you've experienced the newer model, going back feels like downgrading from fiber to dial-up.
It's a reminder of how quickly today's cutting edge becomes yesterday's news in the AI space.
Nvidia's $100 Billion OpenAI Deal Freezes
Perhaps the biggest story of the week, though, happened behind closed doors. Nvidia's up-to-$100 billion investment in OpenAI — announced with great fanfare in September 2025 — is reportedly "on ice," according to The Wall Street Journal.
The deal was designed to cement Nvidia's position as the indispensable compute provider for the AI era. Now, the two sides are "rethinking the future of their partnership," with discussions reportedly shifting to a smaller equity investment of "tens of billions" as part of OpenAI's current funding round.
The freeze signals a potential recalibration of AI's most important supply relationship. Nvidia controls the GPUs that power virtually every major AI model, and OpenAI is the most high-profile consumer of that compute. If the partnership dynamics are shifting, it could have ripple effects across the entire AI industry.
Bottom Line
The AI landscape is maturing rapidly. Anthropic is betting that the future lies in specialized, work-ready tools rather than raw model benchmarks. OpenAI is pruning its model lineup as GPT-5 dominates. And the financial architecture of AI — exemplified by the Nvidia-OpenAI relationship — is being stress-tested. The race is no longer just about who has the best model; it's about who builds the most useful platform.