The model thought for over 5 minutes to produce this. It's not quite photorealistic (some parts are definitely "off"), but this is definitely a significant leap in complexity.
I've been A/B testing the big three (GPT-5, Claude Opus 4, Gemini 3.1) on a real codebase migration this week.
Quick take: Gemini 3.1 Pro's long context is genuinely better now — I fed it a 200k token codebase and it could reference files from the beginning without losing track. That was a real problem in 3.0.
For pure code generation though, Claude still edges it out on following complex multi-step instructions. Gemini tends to take shortcuts when the task has more than ~5 constraints.
The exciting thing is how close they all are. Competition is working exactly as it should.
Comments moved to https://news.ycombinator.com/item?id=47074735, which was posted earlier and is also on the frontpage
Shared this in the other Gemini Pro 3.1 thread (https://news.ycombinator.com/item?id=47074735) but wanted to share it here as well.
I just tested the "generate an SVG of a pelican riding a bicycle" prompt and this is what I got: https://codepen.io/takoid/pen/wBWLOKj
The model thought for over 5 minutes to produce this. It's not quite photorealistic (some parts are definitely "off"), but this is definitely a significant leap in complexity.
Please don't duplicate comments - it makes merging threads a pain!
Letting us know that there's a dupe situation going on is helpful, if you have a minute to email hn@ycombinator.com.
That looks amazing!
But huh, I get... this: https://www.svgviewer.dev/s/1MOs0uyb (on Gemini 3.1 Pro) ... what... is wrong with my Gemini?
I got this : https://www.svgviewer.dev/s/DNq3oGMB
pro with the 20$ subscription , from the gemini.google.com site , initially the request was routed to nano banana,
so I had to change the prompt and had to be very specific about not generating the image directly, here the exact prompt :
"I want the html svg code for that ( not the generated image) , generate an SVG of a pelican riding a bicycle"
I've been A/B testing the big three (GPT-5, Claude Opus 4, Gemini 3.1) on a real codebase migration this week.
Quick take: Gemini 3.1 Pro's long context is genuinely better now — I fed it a 200k token codebase and it could reference files from the beginning without losing track. That was a real problem in 3.0.
For pure code generation though, Claude still edges it out on following complex multi-step instructions. Gemini tends to take shortcuts when the task has more than ~5 constraints.
The exciting thing is how close they all are. Competition is working exactly as it should.
Anyone know if this is available in the Gemini CLI with a subscription (not api key)? Still seeing only Gemini 3 pro preview
Google is known for not giving you an option to give them your money for AI tools. Maybe they will fix this soon.
Is there a way to use Gemini CLI with a $20 plan, like it is with Anthropic or OpenAI CLI agents?
Yes I have the "Google AI Pro" plan from here https://gemini.google/subscriptions/ and you can login to the Gemini CLI with oauth