Gemini 4 Argon

Google, closed weights, released Oct 2026, run at provider defaults through OpenRouter, served by Google AI Studio. 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

61.6

95% interval 54.5–70.9

Rank: 6of 14 models by Index; the paired intervals allow 1 to 11
Cost per page: $0.15Generation only, list price on run day
Second judge: 76.1GPT-5.6 Terra places it seventh of 14. The Claude judge: sixth.
Grade sum: 24.25Seven 1–5 grades, out of 35
Pages judged: 1010 judge passes, 2 builds per brief
Defects per page: 3.1Tap targets under 24px on the phone (8 of 10)

Rank 6 of 14; the data allows anywhere from 1 to 11. Within noise of GPT-6 Astra, Grok 4.7, GPT-6.1 Sol, Claude Sonnet 5.5, GPT-6 Luna, Muse Wonder, Claude Opus 5.5, GPT-5.6 Sol, Claude Fable 5.1 and Claude Fable 5: the gap to each sits inside the noise, so the board can't order them yet. Not clearly behind any model. Clearly ahead of 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.

Hierarchy3.704 of 14
Typography3.357 of 14
Spacing3.157 of 14
Color3.554 of 14
Components3.408 of 14
Composure3.653 of 14
Craft3.455 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.

Landing66.17 of 14
Dashboard54.510 of 14
Settings53.63 of 14
Pricing56.24 of 14
Article77.73 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: 171sMedian over every build, broken ones included
Output tokens: 15.7kMedian per page, reasoning included
Reasoning tokens: 0.0kMedian per page, as the API reports them
Broken pages: 0 of 10Every build came back as a whole page
Criticals: 50%5 of 10 scored pages had a critical issue in their first Rams code review
Built through: OpenRouterServed by Google AI Studio

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: Article83.9
Index on this page, one judge pass. Brief mean 77.7.
- Subsystem meta-value vertical misalignmentModerate
- Drop-cap 'A' detaches from running textModerate
- Uneven vertical rhythm between meta-ledger labels and valuesModerate
Worst: Settings50.0
Index on this page, one judge pass. Brief mean 53.6.
- Redundant page title duplicates sidebar active itemSerious
- Content clipped at right edge throughoutSerious
- Verified badge clipped outside label boundsSerious

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

Head to head

Gemini 4 Argon against every other model: the paired difference in Index points, and whether the 95% interval on it clears zero.

vs GPT-6 Astra-4.8Inside the noise
vs Grok 4.7-4.5Inside the noise
vs GPT-6.1 Sol-3.8Inside the noise
vs Claude Sonnet 5.5-0.2Inside the noise
vs GPT-6 Luna-0.2Inside the noise
vs Muse Wonder+1.1Inside the noise
vs Claude Opus 5.5+1.2Inside the noise
vs GPT-5.6 Sol+2.7Inside the noise
vs Claude Fable 5.1+2.7Inside the noise
vs Claude Fable 5+4.0Inside the noise
vs Gemini 3.8 Flash+6.1Clearly better
vs DeepSeek V4 Pro+7.0Clearly better
vs Claude Opus 4.8+8.8Clearly 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