GPT-6 Luna

OpenAI, closed weights, released Sep 2026, run at provider defaults through OpenRouter, served by OpenAI. It built the same five pages as every model on season 0.5 of the Rams Index, and Rams graded each one at desktop and phone size.

Rams Index, 0 to 100

61.8

95% interval 54.5–69.5

Rank: 5of 14 models by Index; the paired intervals allow 2 to 10
Cost per page: $0.005Generation only, list price on run day
Second judge: 78.2GPT-5.6 Terra places it fifth of 14. The Claude judge: fourth.
Grade sum: 24.30Seven 1–5 grades, out of 35
Pages judged: 1010 judge passes, 2 builds per brief
Defects per page: 5.3Text under 12px (10 of 10)

Rank 5 of 14; the data allows anywhere from 2 to 10. Within noise of Grok 4.7, GPT-6.1 Sol, Claude Sonnet 5.5, Gemini 4 Argon, Muse Wonder, Claude Opus 5.5, GPT-5.6 Sol and Claude Fable 5: the gap to each sits inside the noise, so the board can't order them yet. Clearly behind GPT-6 Astra. Clearly ahead of Claude Fable 5.1, Gemini 3.8 Flash, DeepSeek V4 Pro and Claude Opus 4.8.

Designer check pending · read Nov 6Season 0.5New this seasonProvisional: 1.0 judge passes per page, protocol 2

Grades by dimension

The seven grades behind the Index, 1 to 5, mean of desktop and phone. The tick is the mean across all 14 models.

Hierarchy3.656 of 14
Typography3.503 of 14
Spacing3.157 of 14
Color3.554 of 14
Components3.652 of 14
Composure3.3011 of 14
Craft3.504 of 14

Bars start at 2.8 and end at 4.2, the range the board uses, so small differences show.

Index by brief

The same Index on each of the five briefs, and where the model places on that brief among all 14.

Landing69.63 of 14
Dashboard60.75 of 14
Settings52.74 of 14
Pricing51.86 of 14
Article74.17 of 14

Bars run from 40 to 90.

How it ran

Every model runs at its provider's default settings, through the same harness, with the same budget per page.

Effort: DefaultThe provider’s default. We send no effort or thinking setting
Time per page: 81sMedian over every build, broken ones included
Output tokens: 9.2kMedian per page, reasoning included
Reasoning tokens: 0.8kMedian per page, as the API reports them
Broken pages: 0 of 10Every build came back as a whole page
Criticals: 70%7 of 10 scored pages had a critical issue in their first Rams code review
Built through: OpenRouterServed by OpenAI

Best and worst page

Of its 10 judged pages, the one the judge graded highest and the one it graded lowest, with the judge's top findings on each, most severe first.

Best: Article75.0
Index on this page, one judge pass. Brief mean 74.1.
- Redundant separator characters in bylineModerate
- Code figure caption label misaligned right edgeModerate
- Table 'Change' column lacks semantic color differentiation for neutral rowModerate
Worst: Pricing50.0
Index on this page, one judge pass. Brief mean 51.8.
- Billing toggle save-note vertical misalignmentSerious
- Toggle orphan label misalignedModerate
- Comparison table 'Pro' column header accent isolatedModerate

All 10 of its pages sit beside every other model's in the gallery.

Head to head

GPT-6 Luna against every other model: the paired difference in Index points, and whether the 95% interval on it clears zero.

vs GPT-6 Astra-4.6Clearly behind
vs Grok 4.7-4.3Inside the noise
vs GPT-6.1 Sol-3.6Inside the noise
vs Claude Sonnet 5.50.0Inside the noise
vs Gemini 4 Argon+0.2Inside the noise
vs Muse Wonder+1.2Inside the noise
vs Claude Opus 5.5+1.4Inside the noise
vs GPT-5.6 Sol+2.9Inside the noise
vs Claude Fable 5.1+2.9Clearly better
vs Claude Fable 5+4.2Inside the noise
vs Gemini 3.8 Flash+6.2Clearly better
vs DeepSeek V4 Pro+7.1Clearly better
vs Claude Opus 4.8+8.9Clearly better

The other models

Season 0.5 of the Rams Index, in board order, then the unranked.

Rams Index v0.5 · updated 2026-10-17Judged by Rams eyes · Claude Sonnet 4.6