OpenAI Releases GPT-6 Sol and Luna at Half the API Price

OpenAI released two new models, GPT-6 Sol and GPT-6 Luna, which it describes as bringing much of the strength of its top GPT-6 Astra model into faster, cheaper versions for high-volume work (OpenAI on X). OpenAI says it made caching and inference more efficient and is passing the savings on: API prices for both are 50% lower than the promotional prices of the GPT-5.6 models they follow.

The word promotional matters. Simon Willison notes that GPT-5.6 had a 25% price increase scheduled for November, so the new models cost half of GPT-5.6's discounted rate, and less than half of what GPT-5.6 was due to cost (simonwillison.net). In dollars, Sol goes from $4 to $2 per million input tokens and from $20 to $10 per million output tokens. Luna goes from $0.20 to $0.10 for input and from $1.20 to $0.50 for output, slightly more than half off.

The announcement post gives no benchmark numbers, and its attached launch video could not be reviewed, so on this evidence the claim that the models carry much of Astra's strength is OpenAI's own and unquantified. Artificial Analysis, which ran its own tests on launch day, found both models scoring level with their GPT-5.6 predecessors overall, with gains on some tests and losses on others, and cost per task cut roughly in half by the lower prices (its results). On those measurements the release changes what the work costs more than what the models can do.