AI News · Models ·
OpenAI introduces GPT-6 Sol and Luna
OpenAI introduced GPT-6 Sol and GPT-6 Luna, expanding its GPT-6 model lineup with lower-cost options. Its announcement lists availability in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu users, with Luna also accessible to Free and Go users in the desktop app.
Key points
- OpenAI introduced GPT-6 Sol and Luna as lower-cost additions to its GPT-6 family.
- CNBC reported API prices 50% below GPT-5.6 promotional pricing.
- Sol targets complex workloads such as coding; Luna targets high-volume information extraction and document summarization.
- Availability includes ChatGPT Work and Codex, with Luna also accessible to Free and Go users in the desktop app.
What happened: OpenAI has introduced GPT-6 Sol and GPT-6 Luna, adding lower-cost options to its GPT-6 family of artificial intelligence models. CNBC reported that the releases cut API prices by 50% compared with GPT-5.6 promotional pricing. For business teams, the announcement creates more choice between models intended for complex work and those designed for routine, high-volume tasks. The price comparison is specifically against GPT-5.6 promotional rates, not a reported reduction across every OpenAI product or subscription.
The details: Sol sits one tier below Astra and is intended for more complex workloads, including coding, according to OpenAI’s description cited by CNBC. Luna is aimed at high-volume tasks such as extracting information and summarizing documents. Those examples establish the intended division of work between the two models, but do not establish how well either performs on a particular company’s tasks. Comparative accuracy results, processing speeds and exact per-token API prices were not reported by CNBC, leaving important questions for teams evaluating the trade-offs.
Who it affects: OpenAI’s announcement lists availability in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu users. Luna is also accessible to Free and Go users in the desktop app. That makes the launch relevant both to teams using OpenAI’s applications and to developers paying for model access through an API. The reported API price reduction should be distinguished from application access: a cheaper model API does not, by itself, establish a change in subscription pricing. Changes to subscription charges were not reported.
Background: The launch comes as OpenAI and Anthropic face pressure from cheaper, open-weight competitors, including Alibaba, Moonshot AI and DeepSeek, CNBC reported. Customers are looking for more cost-effective models as they try to control AI spending. Anthropic separately introduced Claude Opus 5.5, describing it as more token-efficient and costing around 40% less to run than Opus 5. These are separate comparisons against each company’s earlier offerings, rather than a direct price comparison between OpenAI and Anthropic.
What to watch: For teams considering a switch, the central question is whether the lower-cost options meet their workload requirements. Testing Sol and Luna on representative workflows before moving work from more expensive alternatives would help assess that fit. OpenAI’s positioning makes coding a relevant test for Sol, and document summarization or information extraction relevant tests for Luna. Lower advertised API prices alone do not resolve the performance question.
Our take
Teams have more options for matching model spending to task requirements. Test both models on representative workflows before moving work from more expensive alternatives.