Anthropic and OpenAI Release Dueling AI Models on the Same Day

Anthropic and OpenAI each released new artificial intelligence models on Thursday. Anthropic unveiled Claude Opus 4.6, which the company says improves performance on office productivity and coding tasks. The model includes an expanded context window that allows it to handle longer documents and more complex projects in one session. OpenAI released GPT-5.3-Codex, a coding-focused model. The company says the model runs faster, uses fewer computing resources and can generate and manage complex software from English instructions. It also comes with a stand-alone Codex desktop app. Executives from both companies appeared on the "TBPN" podcast in back-to-back segments hosted by John Coogan and Jordi Hays. Sam Altman, the OpenAI chief executive, said the field is moving toward workflows in which people manage teams of agents that operate at higher levels of abstraction. Sholto Douglas, an Anthropic researcher, said users have observed that OpenAI models perform better on tough problems while Anthropic models are faster. He added that Anthropic focused on making models better at difficult problems while OpenAI emphasized speed.