OpenAI has expanded its GPT-6 model family with GPT-6 Sol and GPT-6 Luna, introducing two lower-cost reasoning models aimed at coding, agentic workflows and repeatable enterprise tasks. Both models became available on September 22 through the Responses and Chat Completions APIs, alongside access in ChatGPT Work and Codex for eligible paid users.
GPT-6 Sol is positioned as the middle tier between the flagship GPT-6 Astra and the more economical Luna model. It is designed for complex coding and agentic workflows where customers need substantial reasoning capability but do not require the highest-cost model for every request. GPT-6 Luna targets focused, high-volume tasks such as information extraction, document processing, summarisation and other well-defined operational workloads.
For prompts of up to 272,000 input tokens under standard API processing, Sol is priced at $2 per million input tokens, $0.20 per million cached-input tokens and $10 per million output tokens. Luna costs $0.10 per million input tokens, $0.01 per million cached-input tokens and $0.50 per million output tokens. OpenAI said the prices are 50% below the promotional rates for the corresponding GPT-5.6 Sol and Luna models.
Both new models have a 1.05-million-token context window and support outputs of up to 128,000 tokens. They accept text and image inputs and can use web search, file search, code execution, computer-use and function-calling tools through supported interfaces. Reasoning effort can be configured across multiple levels, from none through max, allowing developers to trade latency and cost against the amount of model computation assigned to a task.
Sol’s published standard rate places it at one-fifth of Astra’s $10-per-million-token input price and $50-per-million-token output price. Luna substantially lowers the entry cost for high-volume automated processing. Long-context prompts exceeding 272,000 input tokens are billed at higher rates, while batch and flexible-processing options are available at discounts to standard processing. Regional processing carries an additional charge where offered.
The pricing changes apply directly to Indian software companies, IT-services providers, global capability centres and startups that consume models through usage-based APIs. For large deployments, inference cost influences whether an application can progress from a controlled pilot to continuous production use, particularly for customer support, document operations, software engineering and workflow automation.
OpenAI continues to position Astra as the model for its most demanding reasoning and professional-work tasks. Sol and Luna provide lower-cost alternatives within the same generation rather than replacing that flagship tier. Free and Go users can access Luna through the desktop application, while API customers can call the models using the identifiers gpt-6-sol and gpt-6-luna.
