4 ms·
I tried my "Generate an SVG of a pelican riding a bicycle" prompt against Gemma 3n 7.5GB from Ollama and 15GB for mlx-vlm and got a pleasingly different result
by simonw 1y ago
I tried my "Generate an SVG of a pelican riding a bicycle" prompt against Gemma 3n 7.5GB from Ollama and 15GB for mlx-vlm and got a pleasingly different result for the two quantization sizes: https://simonwillison.net/2025/Jun/26/gemma-3n/ https://simonwillison.net/2025/Jun/26/gemma-3n/
- JohnKemeny 1y agoIs that actually a useful benchmark, or is it just for the laughs? I've never really understood that.
- OtherShrezzing 1y agoFor me, it shows if LLM are generalising from their training data. LLM understand all of the words in the prompt. they understand the spec for svg better than any human. They know what a bird is. They know what a bike is. They know how to draw (and given access to computer-use could probably ace this test). They can plan and execute on those plans. Everything here should be trivial for LLM, but they’re quite poor at it because there’s almost no “how to draw complex shapes in svg” type content in their training set.
- mvdtnz 1y ago[flagged]
- simonw 1y agoScored a whole two upvotes here, my scheme is clearly working great!
- DrammBA 1y agoLeave some upvotes for the rest of us
- Aeolun 1y agoKinda feel like the content is a much better reason to visit than the pelicans. Though I suppose the pelicans are part of the content. I'm quite happy that there's someone with both the time to keep up with all the LLM/AI stuff, that is also good enough at writing amusing stuff that I want to keep reading it.
- mathgeek 1y ago> Kinda feel like the content is a much better reason to visit than the pelicans. That's how the pelicans get ya.
- simonw 1y agoIt was supposed to be a joke. But weirdly it turns out there's a correlation between how good a model is and how good it as at my stupid joke benchmark. I didn't realize quite how strong the correlation was until I put together this talk: https://simonwillison.net/2025/Jun/6/six-months-in-llms/ https://simonwillison.net/2025/Jun/6/six-months-in-llms/
- moritzwarhier 1y agoAlways loved this example, what do you think of ASCII art vs SVG? Since it's not a formal encoding of geometric shapes, it's fundamentally different I guess, but it shares some challenges with the SVG tasks I guess? Correlating phrases/concepts with an encoded visual representation, but without using imagegen, that is. Do you think that "image encoding" is less useful? It's a thing I love to try with various models for fun, too. Talking about illustration-like content, neither text-based ASCII art nor abusing it for rasterization. The results have been interesting, too, but I guess it's less predictable than SVG.
- simonw 1y agoI've had disappointing results with ASCII art so far. Something I really like about SVG is that most models include comments, which give you an idea of what they were trying to do.
- moritzwarhier 1y agoYes, the comments part makes sense, you also included it in the talk (I read the transcript but forgot to mention it in my comment, sorry :) It makes sense, since it works adds associations between descriptions and individual shapes / paths etc., similar to other code.
- throw0363fc53 1y agoNo way *wink*, our DevRel don't push for a good outcome on this one test case to get positive coverage on the top independent blog read by LLM people! https://simonwillison.net/2025/May/20/google-io-pelican/ https://simonwillison.net/2025/May/20/google-io-pelican/
- owebmaster 1y agoI think in 5 years we might have some ultra-realistic pelicans and this benchmark will turn out quite interesting.
- lofaszvanitt 1y agoAnd then the author will try the "Pelican tries to swallow the capybara as-is". And it will fall apart again.
- moritzwarhier 1y agoIt's this part where it gets interesting... how exactly it falls apart :D
- lofaszvanitt 1y agolike a quack
- jerpint 1y agoIt’s been useful though given the authors popularity I suspect it’s only a matter of time new LLMs become “more aware” of it
- dominicrose 1y agoIt's useful because it's SVG so it's different than other image generation methods.
- cedws 1y agoGiven how primitive that image is, what's the point of even having an image model at this size?
- simonw 1y agoThis isn't an image model. It's a text model, but text models can output SVG so you can challenge them to generate a challenging image and see how well they do.
- cedws 1y ago>Multimodal by design: Gemma 3n natively supports image, audio, video, and text inputs and text outputs. But I understood your point, Simon asked it to output SVG (text) instead of a raster image so it's more difficult.
- lastdong 1y agoSo interesting, in the end it (accurately) describes the image. SVG is hard. Made me think we could try introducing a feedback loop until it reaches a close enough representation of what’s asked.