GPT-5.6 Luna
GPT-5.6 Luna
Frontier-family manners at a price that lets you stop counting tokens.
Updated
The read
Luna is the cheap end of the GPT-5.6 line and the most interesting one commercially: it keeps the family's million-token context and structured-output discipline while costing a small fraction of its siblings. For summarising, extraction, routing, and the long tail of small model calls a real product makes, it is hard to argue with. Push it into deep reasoning and the seams show quickly.
Where it shines
- Very cheap for something carrying a frontier family's context window
- Fast enough to sit in a user-facing request path
- Same structured-output and tool-calling behaviour as its larger siblings
Worth knowing first
- Reasoning depth drops off sharply compared with Terra and Sol
- Current price is promotional and has already moved once
Index score
83.3
Provisional. This score is read from the product's public capability, not from the one-prompt rebuild the fully reviewed entries went through, so treat it as a placing rather than a verdict.
- Rank
- 13 of 53
- Builds
- LLMs
- Category
- Small and fast
- Output
- Text and code over an API
- Pricing
- Paid only
- Best for
- The long tail of cheap, high-volume model calls
Scores are our own editorial judgement, weighted craft 55, speed 10, control 15 and value 20.
Compare it with
Also worth a look
Full index- 28Claude Haiku 4.5Claude Haiku 4.5ProvisionalAnthropic's cheap, quick one, for the work that does not need a genius.Small and fastPaid only78.9
- 31Mistral Small 4Mistral Small 4ProvisionalThe best small open model to build a product on, and the licence has no catch.Small and fastFree tier78.1
- 38Grok 4.20Grok 4.20ProvisionalxAI's fast tier: quick, cheap, and priced by how much context you actually use.Small and fastPaid only76.9
