Kimi K3.1 Pro

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

74.1

95% interval 64.0–84.2

Rank: 4of 18 models by Index; the paired intervals allow 1 to 7
Cost per page: $0.32Generation only, list price on run day
Second judge: 83.9GPT-5.6 Terra places it fourth of 18. The Claude judge: sixth.
Grade sum: 27.75Seven 1–5 grades, out of 35
Pages judged: 1010 judge passes, 2 builds per brief
Defects per page: 3.7Text under 12px (10 of 10)

Rank 4 of 18; the data allows anywhere from 1 to 7. Within noise of GLM 5.3, Qwen 3.8 Max, Grok 4.7, GPT-6 Astra, GPT-6.1 Sol and Kimi K3: the gap to each sits inside the noise, so the board can't order them yet. Not clearly behind any model. Clearly ahead of GPT-6 Luna, Gemini 4 Argon, GPT-5.6 Sol, Claude Sonnet 5.5, Muse Wonder, Claude Fable 5.1, Claude Opus 5.5, Claude Fable 5, 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 18 models.

Hierarchy3.609 of 18
Typography3.654 of 18
Spacing3.159 of 18
Color3.655 of 18
Components3.605 of 18
Composure3.752 of 18
Craft3.656 of 18

Bars start at 2.8 and end at 4.4, 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 18.

Landing84.13 of 18
Dashboard66.07 of 18
Settings60.44 of 18
Pricing69.46 of 18
Article90.72 of 18

Bars run from 40 to 100.

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: 694sMedian over every build, broken ones included
Output tokens: 29.1kMedian per page, reasoning included
Reasoning tokens: 15.4kMedian per page, as the API reports them
Broken pages: 0 of 10Every build came back as a whole page
Criticals: 60%6 of 10 scored pages had a critical issue in their first Rams code review
Built through: OpenRouterServed by Moonshot AI

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: Article90.7
Index on this page, one judge pass. Brief mean 90.7.
- Drop-cap 'A' breaks text column alignmentModerate
- Deck paragraph indented relative to headline: misaligned axisModerate
- Pullquote left edge inconsistent with body columnModerate
Worst: Settings58.4
Index on this page, one judge pass. Brief mean 60.4.
- Disabled-state Save button with active status messageSerious
- Inconsistent vertical gap between page sectionsModerate
- 'Save changes' button active despite 'All changes saved' statusModerate