OpenAI has introduced GPT-6 Sol and GPT-6 Luna, two new models positioned as more cost-efficient successors to prior GPT-5.6 variants and Astra. The releases turn an earlier teaser into a concrete product development: Sol and Luna are rolling out through ChatGPT Work, Codex and OpenAI’s API, with OpenAI saying the models can be up to 50% less expensive than comparable GPT-5.6 options.
The announcement arrives in the run-up to OpenAI DevDay 2026, scheduled in San Francisco for September 29. OpenAI has signaled major product and platform announcements around the event, but the immediate practical news is the broader GPT-6 family rollout. Businesses now have a clearer choice between a model designed for demanding professional work and one aimed at high-volume, cost-sensitive workloads.
According to OpenAI’s GPT-6 Sol and Luna announcement, Sol is intended to provide better throughput for more demanding professional tasks at a lower cost, while Luna is intended for workloads where volume and cost efficiency are the priority. Astra remains OpenAI’s top-tier model.
What GPT-6 Sol and Luna change
The most important change is not simply a new model name. OpenAI is presenting Sol and Luna as options for different operating requirements. That matters for teams that need to match model capability and spending to a specific workflow rather than applying the same model to every task.
OpenAI also describes improvements across the GPT-6 family in factuality, coding and computer use. Its announcement references changes to caching and session management that support agent-based workflows. These are relevant details for developers and companies building software that must retain task context, use tools or handle longer-running work.
The rollout spans several OpenAI products and access methods:
- ChatGPT Work and Codex are receiving Sol and Luna for Plus, Pro, Business, Enterprise and Edu users as the rollout progresses.
- Luna is also available to Free and Go users through the desktop app.
-
API users can access the models through the
gpt-6-solandgpt-6-lunaidentifiers. - Astra continues as the highest-tier option in the model family.
OpenAI’s stated pricing direction is particularly notable. Sol and Luna are described as up to 50% cheaper than GPT-5.6 equivalents, although the supplied announcement material does not provide a complete price card or a universal savings figure for every use case. Actual costs will depend on the model selected and the workload a company runs.
| Model | OpenAI positioning | Availability described by OpenAI | Cost context |
|---|---|---|---|
| GPT-6 Sol | Demanding professional tasks and better throughput | Rolling out in ChatGPT Work, Codex and the API | Up to 50% cheaper than GPT-5.6 equivalents |
| GPT-6 Luna | High-volume, cost-sensitive workloads | Rolling out in ChatGPT Work, Codex and the API; also available to Free and Go users through the desktop app | Up to 50% cheaper than GPT-5.6 equivalents |
| Astra | Top-tier model in the family | Referenced as the most capable tier | No comparative pricing detail provided in the announcement material |
A practical model-selection question
For companies using OpenAI through the API or planning new AI-enabled processes, the announcement creates a useful operational question: which work needs the more demanding Sol profile, and which work can run effectively through Luna’s cost-sensitive profile? The answer will depend on the task, required output quality, volume and how the model is incorporated into a product or internal process.
A high-volume workflow could include repeated requests that follow a consistent pattern, while professional work may require more throughput and stronger performance for complex tasks. OpenAI’s positioning does not guarantee that either model is suitable for every use case. Teams should validate quality, reliability and cost against their own workload before standardizing on a model.
The improvements in computer use, caching and session management also point to a broader direction for OpenAI’s platform. These capabilities can matter when an AI system needs to work across multi-step tasks rather than generate a single response. However, OpenAI’s announcement does not specify every implementation detail or every future DevDay product update, so businesses should distinguish the released model capabilities from announcements that may follow later.
OpenAI model updates can also change how customers encounter brands through AI-driven answers. Scalevise helps businesses understand whether they are visible in those responses, identify gaps and prioritize practical improvements without relying on guesswork. Our AI Visibility and GEO Checker provides a focused view of that emerging discovery channel, helping teams connect AI developments to measurable marketing action. Start an AI Visibility scan.
Frequently Asked Questions
What are GPT-6 Sol and GPT-6 Luna?
GPT-6 Sol and GPT-6 Luna are OpenAI models positioned as cost-efficient successors to prior GPT-5.6 variants and Astra. Sol targets demanding professional tasks, while Luna targets high-volume, cost-sensitive workloads.
How much cheaper are GPT-6 Sol and Luna?
OpenAI says Sol and Luna can be up to 50% cheaper than GPT-5.6 equivalents. The announcement material does not provide a single savings figure that applies to every workload.
Where are GPT-6 Sol and Luna available?
OpenAI describes a staged rollout across ChatGPT Work, Codex and API endpoints. API access uses the gpt-6-sol and gpt-6-luna identifiers, and Luna is also accessible to Free and Go users through the desktop app.
What is Astra’s role in the GPT-6 family?
OpenAI describes Astra as the top-tier, most capable model in the family. Sol and Luna add options aimed at professional throughput and cost-sensitive volume respectively.
What should businesses evaluate before using Sol or Luna?
Businesses should test the model against their own tasks, expected output quality, volume and cost requirements. OpenAI’s stated positioning provides a starting point, but suitability depends on the workflow.
Conclusion
GPT-6 Sol and Luna give OpenAI users more explicit choices between professional throughput and high-volume cost efficiency. With availability across ChatGPT Work, Codex and the API, the release is immediately relevant to teams evaluating AI-assisted work and software features. DevDay 2026 may bring further platform news, but the confirmed development is already clear: OpenAI is expanding its model lineup while emphasizing lower-cost access for a broader range of workloads.
Top comments (0)