OpenAI offers distinct models within the GPT-6 family, specifically Astra and Luna, each designed for different operational requirements. Understanding the specific design focus of each model helps developers align their choice with project needs.
GPT-6 Astra and Luna: documented positioning
The GPT-6 Astra model is marketed as a state-of-the-art solution for tasks requiring complex reasoning, coding, document creation, and research. It offers granular control over reasoning efforts, with the low, medium, high, xhigh and max settings. Conversely, the GPT-6 Luna model is designed as an efficient option for focused, high-volume tasks. It also supports various reasoning effort levels, including a 'none' setting, and provides tools for function calling.
Operational Efficiency and Cost Considerations
When selecting between these models, developers often consider cost-efficiency versus performance capabilities. Luna is positioned as the most cost-efficient option in the lineup. Astra, while offering advanced capabilities, includes pricing structures that account for high-input token usage. For instance, prompts exceeding 272,000 input tokens are subject to specific multiplier rates for input, cache, and output. Additionally, both models offer various processing modes, such as Batch and Flex, which are priced at 50% of standard rates, while Fast mode is priced at double the standard rate. The documented long-input multipliers are twice the input and cache rates and 1.5 times the output rate for the full request. These product descriptions and pricing rules are not a measured comparison of code-review accuracy.
Technical Implementation
Both models integrate into development workflows through the OpenAI API. Luna supports the Responses API for built-in tools and function calling. For developers using Chat Completions, function calling with Luna is specifically supported when the reasoning effort is set to 'none'. Users can check the official API model catalog for availability and specific pricing tiers to determine the appropriate model version for their technical infrastructure.
Sources and further reading
GPT-6 Astra Model | OpenAI API
https://developers.openai.com/api/docs/models/gpt-6-astra
GPT-6 Luna Model | OpenAI API
https://developers.openai.com/api/docs/models/gpt-6-luna
Model selection | OpenAI API
https://developers.openai.com/api/docs/guides/model-selection