OpenAI has launched GPT-5.6, introducing new model variants, Luna, Terra, and Sol, to differentiate casual and advanced AI use, while cutting Luna’s costs to maintain its market dominance amid rising competition.
OpenAI has moved to simplify ChatGPT around GPT-5.6, a release the company says is now available across ChatGPT, Codex and its API, with the Sol, Terra and Luna variants split between higher reasoning, balanced performance and lower cost. According to OpenAI’s documentation, Sol is aimed at more demanding work such as coding, research, science and design, while Luna is the fastest and cheapest option and now sits at the centre of the company’s consumer-facing push. Axios reported that OpenAI has also been cutting prices sharply, with Luna becoming far less expensive to run just weeks after launch.
The changes matter because OpenAI is trying to protect its lead in a market where cheaper models are arriving fast, particularly from Chinese rivals. Axios reported that GPT-5.6 Luna has become the default model for all ChatGPT users, while paid subscribers get access to the more capable Sol model for more complex tasks. OpenAI’s help pages say GPT-5.6 is still rolling out gradually, so some users may not see it immediately, and that the Medium, High and Extra High reasoning settings on eligible plans now map to Sol.
That broader reset has also affected the way ChatGPT is positioned for different kinds of use. OpenAI describes Luna as the model for quick, everyday responses, while Sol is pitched for longer, more technical work that may involve multiple steps or deeper reasoning. In effect, the company is separating casual chat from heavier productivity use, a shift that reflects both cost pressure and the growing expectation that AI tools should behave more like general-purpose software than novelty chatbots.
The same week also brought a reminder that not every AI product gets a full launch. A Google AI Studio mobile app briefly appeared in Apple’s App Store listing, but the move appears to have been abandoned before release. The listing suggested Google had been preparing an iPhone app for voice- or text-based prototyping, yet the cancellation underlines how quickly AI vendors are adjusting their product plans as they decide which features belong in a phone app, which belong on the web and which should remain in the browser-based tools they already run.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
