gpt-6-luna
A model series from OpenAI that supports a 1M-token context window and is used for a decision-making API.
Core Facts
OpenAI's GPT series increments by integer version numbers (GPT-3, GPT-4) and uses letter suffixes to distinguish variants, such as "o" representing the optimized version (GPT-4o) and "mini" representing the lightweight version
OpenAI launched the Decisions API at DevDay, powered under the hood by a dedicated gpt-6-luna model.
OpenAI announced that the Decisions API is open for public testing to all developers, powered by GPT-6 Luna and supporting text and image input.
Timeline (last 90 days)
When the user has not set their own default model, LLM 0.37 uses gpt-6-luna as the default model.
During migration, you should keep GPT-4o as a fallback while it still exists, and first migrate a small portion of traffic to observe error rates and quality.
gpt-6-luna reasons at medium effort by default unless reasoning.effort is explicitly set; effort can range from none to max.
gpt-6-luna does not support fine-tuning; OpenAI has restricted new fine-tuning jobs since May 2026, and existing customers will lose the ability to create fine-tuning jobs on January 6, 2027.
On Chat Completions, gpt-6-luna's function calling is available only when reasoning effort is none; using built-in tools requires switching to the Responses API.
One analysis concluded that in short-answer classification scenarios, the cost of the Decisions API is about one-fifth of standard gpt-6-luna pricing.
The Decisions API currently supports only one model, gpt-6-luna, with no fallback model option.
The OpenAI Decisions API (gpt-6-luna) was released on October 6, with an input price of $0.10 / 1M tokens, free output, closed source, and an undisclosed context length.
OpenAI launched the Decisions API at DevDay, powered under the hood by a dedicated gpt-6-luna model.
The author's conclusion: a large number of everyday tasks simply do not require the most expensive model, and Luna matching Opus does not mean it is a drop-in replacement for Opus.
2 more timeline events
All Facts (10)
OpenAI's GPT series increments by integer version numbers (GPT-3, GPT-4) and uses letter suffixes to distinguish variants, such as "o" representing the optimized version (GPT-4o) and "mini" representing the lightweight version
75%VerifiedOpenAI launched the Decisions API at DevDay, powered under the hood by a dedicated gpt-6-luna model.
65%VerifiedOpenAI announced that the Decisions API is open for public testing to all developers, powered by GPT-6 Luna and supporting text and image input.
70%UnverifiedWhen the user has not set their own default model, LLM 0.37 uses gpt-6-luna as the default model.
50%UnverifiedDuring migration, you should keep GPT-4o as a fallback while it still exists, and first migrate a small portion of traffic to observe error rates and quality.
50%Unverifiedgpt-6-luna reasons at medium effort by default unless reasoning.effort is explicitly set; effort can range from none to max.
50%UnverifiedOn Chat Completions, gpt-6-luna's function calling is available only when reasoning effort is none; using built-in tools requires switching to the Responses API.
50%Unverifiedgpt-6-luna does not support fine-tuning; OpenAI has restricted new fine-tuning jobs since May 2026, and existing customers will lose the ability to create fine-tuning jobs on January 6, 2027.
50%UnverifiedThe Decisions API currently supports only one model, gpt-6-luna, with no fallback model option.
50%UnverifiedThe author's conclusion: a large number of everyday tasks simply do not require the most expensive model, and Luna matching Opus does not mean it is a drop-in replacement for Opus.
50%