OpenAI releases GPT-6 Sol and GPT-6 Luna
On September 22, 2026, OpenAI expanded the GPT-6 family with GPT-6 Sol and GPT-6 Luna. OpenAI says the new models bring some of GPT-6 Astra’s advances to faster, more affordable options for everyday work and AI workloads at scale.
GPT-6 Sol: complex coding and agent workflows
OpenAI positions Sol for complex coding and agentic workflows. It accepts text and images, produces text, and supports function calling and Responses API tools including web search, file search and computer use.
Its context window is 1,050,000 tokens and maximum output is 128,000 tokens. Reasoning effort supports none, low, medium, high, xhigh and max.
At Standard rates for prompts of up to 272,000 input tokens, prices per million tokens are $2 for input, $0.20 for cached input and $10 for output. Longer prompts and other processing tiers have different rates.
GPT-6 Luna: focused work at scale
Luna is OpenAI’s most economical GPT-6 model for focused, high-volume tasks. It also accepts text and images and supports function calling and Responses API tools, including web search, file search and computer use.
The context window is 1,050,000 tokens, with a maximum output of 128,000. It offers the same reasoning effort settings: none, low, medium, high, xhigh and max.
At Standard rates for prompts of up to 272,000 input tokens, a million tokens cost $0.10 for input, $0.01 for cached input and $0.50 for output. Long-context and other processing rates differ.
Choosing within the GPT-6 family
- GPT-6 Astra is the family’s most capable option for particularly demanding tasks.
- GPT-6 Sol balances strong capability and cost for complex code and AI agents.
- GPT-6 Luna is designed for focused, high-volume operations at the family’s lowest token price.
Sol and Luna are available through the OpenAI API. OpenAI has also introduced them in Codex and ChatGPT Work; access depends on plan and workspace settings.
The expanded GPT-6 family lets developers match a model to the workload, from complex agent workflows to large volumes of inexpensive requests.