Skip to main content
SuggestedTech
Back to all news

OpenAI

What OpenAI's GPT-5.6 update in ChatGPT means for you

OpenAI has made GPT-5.6 Sol more accurate and given free users unlimited access to GPT-5.6 Luna.

By , Editor-in-Chief · SuggestedTechVerified August 2026

The answer

OpenAI updated ChatGPT's models on 6 August, cutting errors and freeing up Luna access.

OpenAI has updated the models behind ChatGPT, aiming to make answers more accurate and to give free users far more access to the chatbot.

What happened

On 6 August 2026, OpenAI rolled out an updated version of GPT-5.6 Sol, the model used in ChatGPT's main chat product. It is built to give shorter, more direct answers, adjusting how much detail it offers based on the question, using tighter formatting, and pushing back on the user rather than simply agreeing when that would not actually help.

OpenAI tested the update on financial, medical and legal questions. It said responses containing at least one factual error dropped by 68% with GPT-5.6 Sol and by about 62% with GPT-5.6 Luna, compared with the older GPT-5.5 Instant model. OpenAI credits this to better use of the sources the model looks up when answering. These are the company's own tests, not an independent benchmark, so the real-world improvement could differ.

For Plus and Pro subscribers, OpenAI has merged its "fast" and "reasoning" versions of Sol into one model, with a slider letting you choose how much reasoning power to apply to a given question. Free and Go users get a simpler version of this: a new "Think" button they can switch on for harder questions.

Free and Go users also now get GPT-5.6 Luna as their default model, with unlimited text chats, rolling out from 7 August. Some users complained on X after the model picker was moved into ChatGPT's settings menu, making it less immediately visible.

The changes apply only to ChatGPT's consumer chat product. Sol as used in ChatGPT Work and in Codex, OpenAI's coding tool, is unchanged. OpenAI has also strengthened protections for under-18 users, including limits on romantic role-play, according to DataNorth.

What it means for you

If you use ChatGPT for questions involving money, health or legal matters, OpenAI's figures suggest you should see fewer confidently wrong answers, though this depends on independent testing to confirm.

If you're a free or Go user, you can now chat with GPT-5.6 Luna as much as you like, rather than hitting limits. If you pay for Plus or Pro, you no longer need to pick between a fast model and a reasoning model — one model now handles both, and you control how much thinking it does with a slider.

Parents of teenage users should note the tightened limits on romantic role-play for under-18 accounts.

How to try it

The updated GPT-5.6 Sol is already live in ChatGPT for paid users. Free and Go users will see GPT-5.6 Luna become their default model as the rollout continues from 7 August. The model picker now sits inside ChatGPT's settings rather than on the main chat screen.

Sources

← All news