Qwen Image 3.0 Pro • Hacker News Board

Re: Qwen Image 3.0 Pro

#21

Post by giancarlostoro » Wed, Aug 05, 2026, 4:25 PM UTC

NBJack wrote: post #4

If I had a nickel for every model card, repo, or spec posted here that had zero samples of the actual output, I could probably start financing a rack of MI455Xs.

OpenAI does surprisingly well in this regard with their blog posts, and I'm more of an Anthropic fanboy. I wish most tech releases for AI were as well done as OpenAIs they actually showcase what they've built, honestly, if I had to compare it to another company, I wouldn't be surprised if former Apple employees are doing OpenAI's press briefings. Whoever does press briefings at Apple, they are worth their weight in gold.

Re: Qwen Image 3.0 Pro

#22

Post by pixelesque » Wed, Aug 05, 2026, 4:27 PM UTC

phildougherty wrote: post #2

is this going to get released on huggingface or is it cloud/hosted only?

They promised weights for Qwen Image 2 around 6 months ago and there's still nothing, so wouldn't get your hopes up...

Re: Qwen Image 3.0 Pro

#23

Post by saltysalt » Wed, Aug 05, 2026, 4:30 PM UTC

Will be interesting to see if they release the weights for local deployments. I love running Flux2 locally via ComfyUI.

Re: Qwen Image 3.0 Pro

#24

Post by sigmoid10 » Wed, Aug 05, 2026, 4:32 PM UTC

giancarlostoro wrote: post #21

NBJack wrote: post #4

If I had a nickel for every model card, repo, or spec posted here that had zero samples of the actual output, I could probably start financing a rack of MI455Xs.

OpenAI does surprisingly well in this regard with their blog posts, and I'm more of an Anthropic fanboy. I wish most tech releases for AI were as well done as OpenAIs they actually showcase what they've built, honestly, if I had to compare it to another company, I wouldn't be surprised if former Apple employees are doing OpenAI's press briefings. Whoever does press briefings at Apple, they are worth their weight in g…

The last chief communications officer of OpenAI used to work at Apple for 7 years according to her linkedin.

Re: Qwen Image 3.0 Pro

#25

Post by vunderba » Wed, Aug 05, 2026, 4:33 PM UTC

sinak wrote: post #15

Looks like this model is meaningfully less good than gpt-image-2. Arena.ai score is 1263 vs 1380. https://arena.ai/leaderboard/text-to-image

As I’ve said before, I wouldn’t put a lot of stock in Arena’s scoring system. They have MAI Image 2.5 ranked above Gemini Nano Banana Pro, and maybe that’s true on paper but good luck using it. Microsoft’s censorship makes Google feel like the wild west by comparison.

They also have Meta’s Muse Image ranked above NB Pro, which is just patently absurd. In my own GenAI benchmark it only managed a lackluster 7 passes out of 15. Even the open‑weight Ideogram 4 scored higher than that.

For reference, here’s GPT‑Image‑2, NB Pro, and Muse compared:

https://genai-showdown.specr.net/?models=nbp,g2,mi

Re: Qwen Image 3.0 Pro

#26

Post by bayesianbot » Wed, Aug 05, 2026, 4:34 PM UTC

themanmaran wrote: post #12

Pretty annoying user experience trying to test out the model on the official Qwen site. You get an immediate popup with "Security Notice: For security purposes, please add a payment method" that blocks you from clicking anywhere. Honestly if you're launching a new model, you should probably just make it free for the 24 hours after launch so people can test it out. Otherwise just wait till it's on openrouter.

And now you got 90% of the people saying the model sucks as it timed out while serving them

Re: Qwen Image 3.0 Pro

#27

Post by amelius » Wed, Aug 05, 2026, 4:35 PM UTC

NBJack wrote: post #4

If I had a nickel for every model card, repo, or spec posted here that had zero samples of the actual output, I could probably start financing a rack of MI455Xs.

I prefer no samples over cherry picked samples, though.

Re: Qwen Image 3.0 Pro

#28

Post by giancarlostoro » Wed, Aug 05, 2026, 4:41 PM UTC

sigmoid10 wrote: post #24

giancarlostoro wrote: post #21

Earlier quoted context omitted.

OpenAI does surprisingly well in this regard with their blog posts, and I'm more of an Anthropic fanboy. I wish most tech releases for AI were as well done as OpenAIs they actually showcase what they've built, honestly, if I had to compare it to another company, I wouldn't be surprised if former Apple employees are doing OpenAI's press briefings. Whoever does press briefings at Apple, they are worth their weight in g…

The last chief communications officer of OpenAI used to work at Apple for 7 years according to her linkedin.

As I was writing this it dawned on me, and I had a feeling someone would confirm. :)

I am not a Sam Altman fan, but I do give him credit where its due, their press briefings are chef's kiss.

Re: Qwen Image 3.0 Pro

#29

Post by nrightnour » Wed, Aug 05, 2026, 4:46 PM UTC

vunderba wrote: post #25

sinak wrote: post #15

Looks like this model is meaningfully less good than gpt-image-2. Arena.ai score is 1263 vs 1380. https://arena.ai/leaderboard/text-to-image

As I’ve said before, I wouldn’t put a lot of stock in Arena’s scoring system. They have MAI Image 2.5 ranked above Gemini Nano Banana Pro, and maybe that’s true on paper but good luck using it. Microsoft’s censorship makes Google feel like the wild west by comparison. They also have Meta’s Muse Image ranked above NB Pro, which is just patently absurd. In my own GenAI benchmark it only managed a lackluster 7 passes ou…

The photo rankings on this page are so absurd that the only reasonable explanation is that they were judged by an AI.

Re: Qwen Image 3.0 Pro

#30

Post by a11r » Wed, Aug 05, 2026, 4:48 PM UTC

[flagged]