GPT-6 Luna
GPT-6 Luna is OpenAI's efficient GPT-6 model with the API ID gpt-6-luna, a 1,050,000-token context window, 128,000 maximum output, text and image input, text output, and reasoning effort from none through max.
Luna makes long-context reasoning and the current OpenAI tool surface practical for high-volume workloads. Its low base price does not remove the 272K long-prompt multiplier, verification cost, or product-specific access differences.
OpenAI's launch and model page control the API contract. OpenAI's comparative quality and cost claims remain provider-attributed. Community discussions are useful for rollout and usage-limit questions but do not override official availability.
- Process high-volume focused tasks.
- Run low-cost long-context extraction and classification.
- Handle mechanical coding and maintenance with verification.
- Serve as a first-pass route before Sol or Astra.
Base rates are $0.10 input, $0.01 cached input, $0.125 cache write, and $0.50 output per million tokens. Above 272,000 input tokens, 2x input/cache and 1.5x output rates apply to the full request.
GPT-6 can reuse eligible shared prompt prefixes, which matters when many Luna requests carry the same instructions or tool definitions. Use OpenAI's Prompt Caching Dashboard to compare cached and uncached input, inspect diagnostics when a prefix misses, and place an explicit breakpoint after stable content when that fits the workflow. Cache writes have their own price, so compare total request and task cost before claiming savings.
OpenAI began rolling the October GPT-6 Luna into ChatGPT for Free and Go users on October 8, following the October 7 announcement. The September versions in Work and Codex were not changed by this Chat launch. Separate client, workspace and API limits still apply; a plan announcement does not prove every account has received the rollout.
Luna is a candidate for extraction, classification, focused coding, document processing, and first-pass agent work. Escalate tasks when your own evals show the lower-cost route misses requirements; price alone is not a quality guarantee.
Google Cloud’s October 21 announcement names GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna in a Model Garden private preview for selected enterprise customers, with Marketplace billing and Google Cloud APIs. This is not general availability and does not change OpenAI’s direct API IDs, prices or account eligibility. GPT-6 Sol is not the Sol version listed in this preview.
Source confidence
OpenAI
OpenAI Developers
OpenAI
OpenAI Developers
OpenAI
OpenAI
OpenAI
Google Cloud
GPT-6 Luna FAQ
Common questions about GPT-6 Luna.
What is the GPT-6 Luna API model ID?+
Use gpt-6-luna. It supports text and image input, text output, a 1.05M context window, and up to 128K output tokens.
How much does GPT-6 Luna cost?+
Base rates are $0.10 input, $0.01 cached input, $0.125 cache write, and $0.50 output per million tokens, with higher rates above 272K input tokens.
Can Free users access GPT-6 Luna?+
OpenAI documents Free and Go access in the desktop app. That does not imply the same access in Chat, Codex, the API, or every account during rollout.