OpenAI Unifies Paid ChatGPT with GPT‑5.6 Sol and Launches Free Unlimited GPT‑5.6 Luna Chat

Paid users no longer need to choose a model

Previously, ChatGPT users had to decide whether a question could be answered with the Instant model or required the Thinking model. The updated GPT‑5.6 Sol now handles both quick answers and deep reasoning, keeping the response style consistent.

OpenAI announced on X that the fast‑answer and deep‑reasoning experiences for Plus and Pro users are unified under Sol, while Free/Go users receive unlimited text chat with GPT‑5.6 Luna .

A slider takes over fast‑answer and deep‑thinking

Plus and Pro users see a new “thinking amount” slider on web, mobile, and desktop. Users can still choose “faster” or “think more,” but the control shifts from selecting a model name to adjusting inference compute.

OpenAI provided internal benchmark numbers: in fact‑checking prompts for finance, medical, and legal domains, answers containing at least one factual error were reduced by 68 % with Sol and 62 % with Luna compared to GPT‑5.5 Instant. These figures come from OpenAI’s own evaluation and are not third‑party benchmarks.

Free tier now gets Luna

Free and Go users default to GPT‑5.6 Luna with unlimited text‑only conversation. A “Think” button lets Luna spend more inference time on difficult queries.

Unlimited usage has three limits: only text chat is unlimited; file uploads, images, and other tools remain quota‑limited; the “Think” feature is subject to anti‑abuse rules; and free users receive Luna, not the paid‑tier Sol.

Chat experience simplified, but routing becomes less transparent

The update affects only the Chat experience; the Work and Codex models continue using the existing Sol version, so “Sol updated” does not apply to all OpenAI products.

For ordinary users, the main benefit is eliminating a model‑selection step: ask a question, then use the slider or Think button to indicate how much reasoning is needed. The trade‑off is that model routing is hidden—different users still receive varying capabilities, speed, and tool limits under the same ChatGPT brand.

For paid users, the unified Sol experience reduces mode‑switching, but the claimed 68 % error‑rate reduction still needs real‑world validation. Early community reactions focus on whether Luna will be rate‑limited, how it compares to Sol on complex tasks, and whether higher reasoning settings noticeably slow responses.

From a competitive standpoint, OpenAI’s pressure will fall on two fronts: free text chat becoming a baseline offering, and paid value needing to rely on stronger models, adjustable inference, and expanded tool quotas. Future user decisions may shift from remembering model names to asking whether the system can automatically allocate the appropriate capability and clearly state its limits.

GPT‑5.6 Luna offers free unlimited text chat; GPT‑5.6 Sol unifies paid ChatGPT fast‑answer and deep‑thinking. Free users do not get Sol, and Work/Codex remain unchanged.