OpenAI · Status at source date: Official announcement
GPT-6 Sol and Luna broaden the options for everyday work
OpenAI’s announcement makes Sol and Luna available through ChatGPT Work, Codex, and the API, with access varying by plan; it also distinguishes these surfaces from Chat. For marketers, compare the models on repeatable analysis and production tasks, measuring factual quality and review effort alongside cost.
What changed
The published benchmarks are vendor evaluations, not a guarantee of performance on a client workflow.
Build a small repeatable set of marketing tasks with known review criteria. Compare factual accuracy, instruction following and time to an acceptable result, including human correction effort. Confirm the product surface and plan where each model is available rather than assuming that ChatGPT, Codex and the API have identical access.
The source presents Sol and Luna as options for everyday work within the GPT-6 family. Product availability and the way a model is exposed can differ between ChatGPT, Codex and the API, so a team should identify the actual surface used in its process. A model name in an announcement is not enough to establish the tools, context, permissions or usage limits available in a particular account.
Use tasks with a reviewable answer or quality rubric, such as explaining a known data discrepancy or drafting a brief from a fixed set of facts. Record mistakes and omissions as well as speed. For code or analysis that will be reused, check reproducibility and the effort needed to validate the output. This produces an informed selection for a specific task rather than a general claim that a cheaper or faster model can replace every step of the existing process.
Why it matters for advertisers
More model options make task-level evaluation more useful than choosing a single model by brand or headline capability. Quality, latency and cost can matter differently for an initial draft, a data explanation or code used in an operational workflow.
What to check next
- Compare models on a repeatable marketing task using quality, latency and cost together.
Sources & contributor credit
- Original source · OpenAI announcement
Original announcement or reported observation; scope retained in the summary.
Original creator unverified
The names below are source credits. The original creator and first-reporting priority have not yet been independently verified.
Attribution evidence and limitations
Full attribution review pending.
A source link alone does not establish original authorship.
Original image and video creators have not yet been verified.
- Published on this site
This update reflects the dated source reporting. Availability may have changed. Further coverage of this same development will be added to this page.
