跳到正文
Simon Willison 博客· Simon Willison·· 2026-04-17AI 评分49

我笔记本上的 Qwen3.6-35B-A3B 画得比 Claude Opus 4.7 更好的鹈鹕

Qwen3.6-35B-A3B on my laptop drew me a better pelican than Claude Opus 4.7

AI 导读

阿里巴巴发布 Qwen3.6-35B-A3B 后,科技博主 Simon Willison 用其自创的"骑自行车鹈鹕"SVG 生成测试对比了该模型与 Anthropic 新发布的 Claude Opus 4.7。结果显示,仅需 20.9GB 显存的量化版 Qwen3.6-35B-A3B 在 MacBook Pro M5 本地运行时,生成的鹈鹕图像在自行车结构、背景细节和鸟类特征上均优于 Opus 4.7。在另一项"骑独轮车火烈鸟"测试中,Qwen 同样胜出。作者指出,尽管这并不意味着该模型整体性能超越 Claude,但在特定图像生成任务上,本地运行的开源模型已能击败顶级闭源对手。

正文 · 原文

For anyone who has been (inadvisably) taking my pelican riding a bicycle benchmark seriously as a robust way to test models, here are pelicans from this morning’s two big model releases—Qwen3.6-35B-A3B from Alibaba and Claude Opus 4.7 from Anthropic.

Here’s the Qwen 3.6 pelican, generated using this 20.9GB Qwen3.6-35B-A3B-UD-Q4_K_S.gguf quantized model by Unsloth, running on my MacBook Pro M5 via LM Studio (and the llm-lmstudio plugin)—:

And here’s one I got from Anthropic’s brand new Claude Opus 4.7 ():

I’m giving this one to Qwen 3.6. Opus managed to mess up the bicycle frame!

I tried Opus a second time passing thinking_level: max. It didn’t do much better ():

I don’t think Qwen are cheating

A lot of people are convinced that the labs train for my stupid benchmark. I don’t think they do, but honestly this result did give me a little glint of suspicion. So I’m burning one of my secret backup tests—here’s what I got from Qwen3.6-35B-A3B and Opus 4.7 for “Generate an SVG of a flamingo riding a unicycle”:

Qwen3.6-35B-A3B
()
The unicycle spokes are a too long. The pelican has sunglasses, a bowtie and appears to be smoking a cigarette. It has two heart emoji surrounding the caption Flamingo on a Unicycle. It has a lot of charisma.
Opus 4.7
()
The unicycle has a black wheel. The flamingo is a competent if slightly dull vector illustration of a flamingo. It has no flair.

I’m giving this one to Qwen too, partly for the excellent <!-- Sunglasses on flamingo! --> SVG comment.

What can we learn from this?

The pelican benchmark has always been meant as a joke—it’s mainly a statement on how obtuse and absurd the task of comparing these models is.

The weird thing about that joke is that, for the most part, there has been a direct correlation between the quality of the pelicans produced and the general usefulness of the models. Those first pelicans from October 2024 were junk. The more recent entries have generally been much, much better—to the point that Gemini 3.1 Pro produces illustrations you could actually use somewhere, provided you had a pressing need to illustrate a pelican riding a bicycle.

Today, even that loose connection to utility has been broken. I have enormous respect for Qwen, but I very much doubt that a 21GB quantized version of their latest model is more powerful or useful than Anthropic’s latest proprietary release.

If the thing you need is an SVG illustration of a pelican riding a bicycle though, right now Qwen3.6-35B-A3B running on a laptop is a better bet than Opus 4.7!

16th April 2026

More recent articles

This is Qwen3.6-35B-A3B on my laptop drew me a better pelican than Claude Opus 4.7 by Simon Willison, posted on 16th April 2026.

local-llms

qwen

lm-studio

来源:Simon Willison 博客 · simonwillison.net