OpenAI's GPT-6 Sol and Luna cost half as much to run
By Brett Gavaghan, Founder of Adaptd · · Updated
OpenAI released two new GPT-6 models on 22 September. They are in ChatGPT for paid plans and cost 50% less on the API than the GPT-5.6 models they follow.
The short version
- OpenAI released two new models, GPT-6 Sol and GPT-6 Luna, on Tuesday 22 September.
- Both are in ChatGPT on paid plans and cost 50% less on the API than GPT-5.6 at its promotional pricing.
- Test a handful of real jobs on the new model before moving a working process across.
OpenAI released two new models on Tuesday 22 September, GPT-6 Sol and GPT-6 Luna.
What are GPT-6 Sol and GPT-6 Luna?
GPT-6 Sol and GPT-6 Luna are both built on the work behind GPT-6 Astra, OpenAI's flagship, and OpenAI describes them as two different trade-offs between capability and cost.
OpenAI also says both improve on their GPT-5.6 counterparts in alignment, and Plus and Pro subscribers, along with Business accounts, got a one-off usage reset added to their accounts to use when they choose.
Who can use them, and what do they cost?
In ChatGPT, Sol and Luna are available to Plus, Pro, Business, Enterprise and Edu users, and in Codex. Free and Go users can try Luna, but only in the desktop app. On the API, OpenAI says both models cost 50% less than GPT-5.6 at its promotional pricing.
For a midsize business, the price is the part to notice. If you pay for a tool or an automation that calls OpenAI's API, the same job can now run for about half the model cost. For teams on ChatGPT Business, you now have two new models to pick from, so it is worth knowing which one people are using for what.
Should your business move to the new models?
Before you move a working process to a new model, test it on your own tasks. Pick a handful of real jobs your team does, such as customer emails or report summaries, and compare the output side by side. If an automation is set to a specific older model name, it will stay on that model until someone changes it.
Cheaper models make it easier to try AI on work that did not justify the cost before. Test first, then switch.
Wondering what this means for your business?
That's exactly what the discovery call is for.