What is the best AI model for 3D scenes?

The model at the top of this board, ranked by people's blind votes between 3D scenes built in code in Design Arena. It updates every week, so the top of the table is the leader today.

How does Design Arena rank AI models?

People compare two designs made from the same prompt without seeing which model made them, and vote for the better one. Design Arena turns those votes into a rating with the Bradley-Terry model.

What does win rate mean?

It's the share of head-to-head votes a model won.

What is Design Arena?

It's a site where people vote blind between designs that AI models made from the same prompt. It calls itself the largest crowdsourced benchmark for AI-generated design.

Can I trust blind votes?

Blind votes remove brand bias, since voters don't know which model made which design. They measure which result people prefer, so they're a guide to how a design looks and works, not to how clean its code is.

Why is the price shown next to the rating?

A model that wins a few more votes can cost many times more. The chart of rating against price shows which models give you the most for what they cost.

How often is this board updated?

Every week, from OpenRouter's public model list. The date under the table shows when the board was last checked.

Where do these numbers come from?

Every model and price comes from OpenRouter's public model list. The scores are Artificial Analysis's and Design Arena's, carried in that same list.

Can I download the data?

Yes: every board has a CSV of its models, next to its chart.

What does open weights mean?

The maker publishes the model's weights, so anyone can download the model and run it on their own hardware.

Why do some models have no score?

Not every model on OpenRouter has been tested on every measure yet. A board lists only the models with a result on its own measure.

Best AI model for 3D scenes

Find the best AI model for 3D scenes among 101 AI models, ranked by Design Arena, where people vote blind between two 3D scenes built in code made from the same prompt.

Updated Oct 1, 2026

Rank	Model
1	Claude Opus 5.5Anthropic	1,457Lead	73.2%Lead	$8	Sept 22, 2026
2	Muse Spark 1.3Meta	1,413#2	64.4%	$2	Sept 2, 2026
3	Claude Fable 5.1Anthropic	1,408#3	67.8%#3	$20	Sept 1, 2026
4	Kimi K3Open weightsMoonshot AI	1,406	68.0%#2	$3.00	Jul 16, 2026
5	GLM 5.3Open weightsZ.ai	1,372	66.0%	$1.01	Aug 18, 2026
6	Claude Opus 5Anthropic	1,347	61.0%	$10	Jul 24, 2026
7	MiMo-V2.6-ProOpen weightsXiaomi	1,342	57.8%	$0.54	Sept 21, 2026
8	GLM 5.1Open weightsZ.ai	1,336	62.6%	$1.48	Apr 7, 2026
9	Claude Fable 5Anthropic	1,334	61.8%	$20	Jun 9, 2026
10	GLM 5.3 FlashOpen weightsZ.ai	1,334	56.8%	$0.24	Aug 26, 2026
11	Gemini 3.7 FlashGoogle	1,326	54.8%	$1.50	Aug 13, 2026
12	Muse Spark 1.2Meta	1,322	53.1%	$2	Aug 5, 2026
13	Gemini 3.8 FlashGoogle	1,312	50.7%	$1.50	Sept 2, 2026
14	GLM 5.2Open weightsZ.ai	1,305	53.6%	$1.31	Jun 16, 2026
15	Grok 4.6xAI	1,292	52.5%	$3	Aug 12, 2026
16	Claude Opus 4.6Anthropic	1,289	62.5%	$10	Feb 4, 2026
17	Gemini 3.6 FlashGoogle	1,288	53.2%	$1.50	Jul 21, 2026
18	Kimi K2.6Open weightsMoonshot AI	1,283	57.6%	$0.78	Apr 20, 2026
19	Qwen3.7 MaxQwen	1,281	54.9%	$2.21	May 21, 2026
20	Claude Sonnet 5Anthropic	1,276	54.4%	$4	Jun 30, 2026
21	Muse Spark 1.1Meta	1,273	49.4%	$2	Jul 16, 2026
22	Grok 4.5xAI	1,273	50.3%	$3	Jul 8, 2026
23	DeepSeek V4 Pro 0423Open weightsDeepSeek	1,269	56.8%	$0.26	Apr 24, 2026
24	GLM 5 TurboZ.ai	1,264	56.0%	$1.90	Mar 15, 2026
25	MiMo-V2.5-ProOpen weightsXiaomi	1,261	54.5%	$0.54	Apr 22, 2026

Source: Design Arena's blind votes and OpenRouter's prices, checked Oct 1, 2026

Blind votes · Design Arena

Top 10: 3D scenes

Oct 1, 2026

11,457Claude Opus 5.5
21,413Muse Spark 1.3
31,408Claude Fable 5.1
41,406Kimi K3
51,372GLM 5.3
61,347Claude Opus 5
71,342MiMo-V2.6-Pro
81,336GLM 5.1
91,334Claude Fable 5
101,334GLM 5.3 Flash

Source: Design Arena's blind votes and OpenRouter's prices, checked Oct 1, 2026

How to read this chart

The bars show this board's 10 best models, best first, with the top 3 in the brand's blue. The bars start just below the lowest score, so small gaps show: read the numbers on the bars. Hover a bar for its maker, or switch to Price to see what each one costs.

Comparison details

Ratings come from people's blind votes between two results of the same prompt on Design Arena, and prices are OpenRouter's blended list prices per million tokens, checked Oct 1, 2026. A rating compares models on this board only.

Every leaderboard

Each board has its own page with every model, a chart and a CSV.

LLM leaderboard

Claude Opus 5.5 leads, from Anthropic

Best AI model for coding

Claude Fable 5.1 leads, from Anthropic

Best AI model for agents

Claude Fable 5.1 leads, from Anthropic

AI model pricing

Cheapest now: Mistral Nemo

Best AI model for writing

Best AI model for website design

Muse Spark 1.3 leads, from Meta

Best AI model for game development

Claude Opus 5.5 leads, from Anthropic

Best AI model for UI design

Claude Opus 5.5 leads, from Anthropic

Best AI model for data visualization

Claude Opus 5.5 leads, from Anthropic

Best AI model for SVG graphics

Claude Opus 5.5 leads, from Anthropic

Best AI model for building mobile apps

Claude Opus 5 leads, from Anthropic

Best AI model for building full-stack apps

Claude Fable 5.1 leads, from Anthropic

Best AI image model

GPT Image 2.5 Sunburst (max) leads, from OpenAI

Best AI image editing model

GPT Image 2.5 Sunburst (max) leads, from OpenAI

Best AI video model

Wan 3.0 leads, from Alibaba

Best image to video AI model

Gemini Omni Flash leads, from Google

Best AI voice model

Eleven v4 leads, from ElevenLabs