OpenAI Expands GPT-6 Lineup With Sol and Luna Models

OpenAI has announced the release of GPT-6 Sol and Luna, expanding the GPT-6 model family following the earlier introduction of GPT-6 Astra. The new models are designed to bring high-level capabilities in professional work, factuality, coding, and computer use to more accessible price points.

API pricing for the new models has been reduced by 50% compared with their GPT-5.6 promotional pricing. For GPT-6 Sol, input pricing is set at $2 per million tokens and output pricing is $10 per million tokens, down from $4 and $20. For GPT-6 Luna, input pricing is $0.10 per million tokens and output pricing is $0.50 per million tokens, down from $0.20 and $1.20.

In benchmark evaluations, GPT-6 Sol at xhigh effort achieved a 33.2% score on AutomationBench at a cost of $0.27 per task. On Agents' Last Exam, GPT-6 Sol at max effort scored 56.4%. For software engineering evaluations on DeepSWE v1.1, GPT-6 Sol at max effort scored 68.8%, while GPT-6 Luna at max effort scored 66.6%. On OSWorld 2.0 offline, GPT-6 Sol at xhigh effort reached a score of 60.5%.

The models also incorporate updated communication styles intended to reduce jargon and provide clearer answers in technical and coding interactions.