gpt-5.6-luna-max: LiveBench scores and how to run it with BrainOutput
gpt-5.6-luna-max is a model from OpenAI. In the LiveBench release of 2026-06-25 (fetched 2026-10-03) it ranks #45 of 66, with a global score of 73.56.
What is gpt-5.6-luna-max?
gpt-5.6-luna-max is listed on LiveBench, a public, continuously refreshed benchmark that scores models on seven categories: reasoning, coding, agentic coding, mathematics, data analysis, language and instruction following. Its strongest category in this snapshot: coding (#9 of 66).
Its weights are not published: it is available only through its vendor's service. With BrainOutput it stays a cloud option for non-sensitive work, when your policy allows it.
The model at a glance
| Vendor | OpenAI |
|---|---|
| Weights | Proprietary (weights not published) |
| Licence | proprietary (API only) |
| Release date (per LiveBench) | — |
| LiveBench identifier | gpt-5.6-luna-max |
LiveBench scores, dated
Score and rank out of 66 models, per category, in the LiveBench release of 2026-06-25 fetched 2026-10-03.
LiveBench, release 2026-06-25 · livebench.ai · 66 models · fetched 2026-10-03
Overall rank in this LiveBench snapshot.
Global score: the mean of the seven categories.
Gap to #1, Claude Fable 5.1 Max Effort (83.41).
| Category | Score | Rank |
|---|---|---|
| Reasoning | 85.64 | #36 / 66 |
| Coding | 82.91 | #9 / 66 |
| Agentic coding | 48.43 | #46 / 66 |
| Mathematics | 87.20 | #45 / 66 |
| Data analysis | 78.03 | #28 / 66 |
| Language | 72.57 | #58 / 66 |
| Instruction following | 60.12 | #58 / 66 |
| Global (mean of the 7 categories) | 73.56 | #45 / 66 |
Method note: the global score is the mean of the seven category averages, computed by us from LiveBench's published file; it may differ by a few decimals from the site's own display. Equal scores keep LiveBench's file order. Benchmarks measure benchmark tasks — during an AI diagnostic we re-run the comparison on your own workloads.
Running gpt-5.6-luna-max with BrainOutput
BrainOutput routes each task to a model by confidentiality: sensitive work stays on a local or private model, and everything else goes to a cloud model only when your policy allows it.
Locally, on your hardware
Not possible
Not possible: the weights are not published. Sensitive work goes to an open-weight model running locally instead.
In a private cloud
Not possible
Not possible as private inference: only the vendor serves this model.
With your own key (BYOK)
Available
Plug in your own OpenAI API key: BrainOutput routes non-sensitive work to it when your policy allows; sensitive work never leaves your perimeter.
Frequently asked questions
Is gpt-5.6-luna-max an open-weight model?
No: its weights are not published. The highest-ranked open-weight model in this snapshot is DeepSeek V4.1 Flash Max Effort (#9, 81.11).
Can we use gpt-5.6-luna-max with BrainOutput?
Yes, through your own OpenAI API key (BYOK), for non-sensitive work and only when your policy allows it. Sensitive work stays on an open-weight model running locally or in a private cloud.
Where do these scores come from?
From LiveBench (livebench.ai), release 2026-06-25, file fetched 2026-10-03. The global score is the mean of the seven categories, computed by us. The ranking changes monthly; we re-verify at diagnostic time.
Find out whether gpt-5.6-luna-max fits your work
An AI diagnostic maps your processes, classes what is sensitive, and benchmarks current models against your real tasks — dated numbers, yours this time.
Sources
- LiveBench — leaderboard, release 2026-06-25 — accessed 2026-10-03
- gpt-5.6-luna-max — vendor and licence — accessed 2026-10-04