Philip Kiely, an engineering expert at the AI model inference infrastructure company Baseten, tweeted that enterprise teams can now natively use open models such as GLM-5.3 Flash and Kimi K3 in Codex, and related call expenses can be directly included in their OpenAI commitment amount. In response, Tibo Sottiaux, OpenAI's core product director and Codex core developer, retweeted the tweet and commented: “I'm proud of this. Openness is the way out.”

Philip Kiely, an engineering expert at the AI model inference infrastructure company Baseten, tweeted that enterprise teams can now natively use open models such as GLM-5.3 Flash and Kimi K3 in Codex, and related call expenses can be directly included in their OpenAI commitment amount. In response, Tibo Sottiaux, OpenAI's core product director and Codex core developer, retweeted the tweet and commented: “I'm proud of this. Openness is the way out.”