The most interesting part to me is the bike. I don't think any of them would actually work, but the mistakes feel somewhat human.
Re:GPT-6 Astra
by irthomasthomas on Friday September 04, 2026 @07:00PM
(#49571197)
Technically on par with gemini flash 3.8, but I give Astra more points for style, and for breaking from the pack by not adding the headgear, or the fish in a basket.
It's a different prompt. It says his prompt was "Generate an SVG of a pelican riding a bicycle" but one I saw earlier this week specifically talked about the blue helmet and fish in a basket.
Re:GPT-6 Astra
by irthomasthomas on Friday September 04, 2026 @08:56PM
(#49571990)
I don't think so: https://news.ycombinator.com/item?id=49538217
It looks like I misinterpreted the bottom of this page[0] where it says:
> A seaside scene: a white pelican in a blue helmet pedals a red bicycle along the beach, one wing gripping the handlebar, its big orange beak and pouch out front, webbed feet on the pedals, and a fish tucked in the front basket for later.
I thought that was the prompt from how it was worded. I guess it's part of the response.
[0] https://tools.simonwillison.net/markdown-svg-renderer?url=ht... scroll to the bottom
I think the Gemini 3.8 Flash ones the other day were superior pelicans. Particularly the murderous one who was going to squish his tiny cousin.
Interesting that terra xhigh effort is cycling right left, in opposition to the prevailing pelican left to right direction.
Interesting that the composition is stable across different runs. There is much more variety in other models.
Especially interesting that once Astra gets to high, xhigh, max it uses the same aesthetic.
Pretty nice pelicans!
63 cents is quite cheap isn't it? If you compare vs Fable5.1 Max @ a whooping $3.30
Even the "low" pelican is better than most other models.
Re:GPT-6 Astra
by matt_heimer on Friday September 04, 2026 @06:34PM
(#49571013)
It's subjective but I think only gpt-5.6-sol XHIGH is comparable to gpt-6-astra LOW in quality. And it's 24.11 cents vs 9.55 for astra low.
Something I've noticed especially with fable at work is it seems to be smarter, so it takes fewer tokens and burns less of my usage than smaller models would for similar tasks.
The "max" pelican looks very serious, almost as if it's determined to win the race!
Yes! And do you get the impression, like I do, that model effort and pelican effort seem correlated?
Yes that's how the vectors work.
This overall issue has resulted in hilarious missteps in the past, including grok going off the rails and claiming to be mechahitler.
At this point in time, we should be able to switch to 3D pelicans in a navigable environment.
Re:GPT-6 Astra
by eh_why_not on Saturday September 05, 2026 @06:28AM
(#49575117)
FWIW, they had a similar-looking pelican (with the red bandana on the neck and all) in their release video:
Love it! You should do a big table comparing pelicans for all the models you tested across all the big companies!
The "max" one at the bottom took 4 minutes 2 seconds and cost 63.206 cents.
For comparison, here those new Astra pelicans are in a grid with the GPT-5.6 pelicans: https://static.simonwillison.net/static/2026/gpt-6-and-5.6-p...