This week, the conflict between OpenAI and Anthropic escalated.
The two companies released competing new AI models Thursday and appeared on back-to-back podcasts on “TBPN.”
Anthropic announced Claude Opus 4.6 on Thursday. It’s an upgraded model that the company claims improves office productivity and performance for coding tasks, with an expanded “context window” that allows users to handle longer documents and more complex projects in a single session.
Meanwhile, OpenAI has fought back with its own new coding-focused model called GPT-5.3-Codex. The company says the model runs faster, uses fewer computing resources, and can generate and manage complex software from English-language instructions. The new version also comes with a standalone Codex desktop app.
OpenAI CEO Sam Altman and Sholto Douglas, one of Anthropic’s principal researchers, appeared on the TBPN podcast and had a back-to-back chat with hosts John Coogan and Geordie Hayes.
“I think we’re going to move toward a workflow where a lot of people feel like they’re just managing a team of agents,” Altman said. “And as agents improve, they will continue to operate at higher and higher levels of abstraction.”
Appearing in a later timeslot, Douglas told Coogan and Hayes that users had been comparing the previous Anthropic model to the OpenAI model and noticed some key differences.
“The OpenAI model was a little bit better at working really, really, really hard on hard problems, but the Anthropic model was much faster,” Douglas said.
“And they worked on speed and we worked on making the models much better at really, really hard problems,” Douglas added of Opus 4.6.
The latest release is part of a long-running competition between Anthropic and OpenAI that dates back to 2021. At this time, a group of OpenAI researchers left to found Anthropic with the goal of developing safer and more controlled AI systems.
A big week for Anthropic
This week, Anthropic’s announcement of an industry-specific plug-in caused the stock market to fall as Wall Street worried about the impact of AI on software.
Anthropic also took a sophisticated shot at OpenAI in a series of ads released this week, including one that will air during the Super Bowl.
The ad features an unnamed humanized AI dropping an ad in the middle of giving advice, with the promise that its model, Claude, won’t see any ads.
In January, OpenAI announced that ads would be served on ChatGPT for users of the free version.
Altman later fired back, calling Anthropic “disingenuous” and defending ChatGPT as a product that brings AI “to the billions of people who can’t afford a subscription.” He also revealed that ads will be “clearly labeled” to distinguish them from answers to chatbot questions.
“We’re not stupid. We respect our users, and we understand that if you do what’s portrayed in the ad, people will naturally stop using the product,” Altman said Thursday on the TBPN podcast.
“Our first principle with advertising is to not put anything in the LLM stream,” Altman added. “It’s going to feel like an outrageous dystopia, like a bad sci-fi movie.”
