Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Qwen3: Think deeper, act faster

qwenlm.github.io

Post reply on HN
Print view View original

261–270 of 412 posts

Re: Qwen3: Think deeper, act faster

#261

Post by stavros »

oscord wrote: post #231
stavros wrote: post #134

Earlier quoted context omitted.

I don't really want it added to the training set, but eh. Here you go: > Assume I have a 3D printer that's currently printing, and I pause the print. What expends more energy, keeping the hotend at some temperature above room temperature and heating it up the rest of the way when I want to use it, or turning it completely off and then heat it all the way when I need it? Is there an amount of time beyond which the ans…

Qwen3-32b did it pretty accurately it seems. Calculated heat loss over time going to ambient temp, offered to keep it at standby 100C for short breaks under 10 minutes. Shut down completely for longer breaks.

The correct answer is that it's always better to turn it off, though.

Re: Qwen3: Think deeper, act faster

#262

Post by phh »

hbbio wrote: post #246
simonw wrote: post #207

As is now traditional for new LLM releases, I used Qwen 3 (32B, run via Ollama on a Mac) to summarize this Hacker News conversation about itself - run at the point when it hit 112 comments. The results were kind of fascinating, because it appeared to confuse my system prompt telling it to summarize the conversation with the various questions asked in the post itself, which it tried to answer. I don't think it did a g…

I also have a benchmark that I'm using for my nanoagent[1] controllers. Qwen3 is impressive in some aspects but it thinks too much! Qwen3-0.6b is showing even better performance than Llama 3.2 3b... but it is 6x slower. The results are similar to Gemma3 4b, but the latter is 5x faster on Apple M3 hardware. So maybe, the utility is to run better models in cases where memory is the limiting factor, such as Nvidia GPUs?…

What's cool with those models is that you can tweak the thinking process, all the way down to "no thinking". It's maybe not available in your inference engine though

Re: Qwen3: Think deeper, act faster

#263

Post by hbbio »

phh wrote: post #262
hbbio wrote: post #246

Earlier quoted context omitted.

I also have a benchmark that I'm using for my nanoagent[1] controllers. Qwen3 is impressive in some aspects but it thinks too much! Qwen3-0.6b is showing even better performance than Llama 3.2 3b... but it is 6x slower. The results are similar to Gemma3 4b, but the latter is 5x faster on Apple M3 hardware. So maybe, the utility is to run better models in cases where memory is the limiting factor, such as Nvidia GPUs?…

What's cool with those models is that you can tweak the thinking process, all the way down to "no thinking". It's maybe not available in your inference engine though

Feel free to add a PR :)

What is the parameter?

Re: Qwen3: Think deeper, act faster

#264

Post by aitchnyu »

estsauver wrote: post #244
kristianp wrote: post #230

Earlier quoted context omitted.

It would be interesting to try, but for the Aider benchmark, the dense 32B model scores 50.2 and the 30B-A3B doesn't publish the Aider benchmark, so it may be poor.

Is that Qwen 2.5 or Qwen 3? I don't see a qwen 3 on the aider benchmark here yet: https://aider.chat/docs/leaderboards/

As a human who asks AI to edit upto 50 SLOC at a time, is there value in models which score less than 50%? Im using the `gemini-2.0-flash-001` though.

Re: Qwen3: Think deeper, act faster

#265

Post by aitchnyu »

daemonologist wrote: post #32
thierrydamiba wrote: post #21

Earlier quoted context omitted.

How do people typically do napkin math to figure out if their machine can “handle” a model?

The ultra-simplified napkin math is 1 GB (V)RAM per 1 billion parameters, at a 4-5 bit-per-weight quantization. This usually gives most of the performance of the full size model and leaves a little bit of room for context, although not necessarily the full supported size.

How much memory would correspond to a 100000 and a million tokens?

Re: Qwen3: Think deeper, act faster

#266

Post by rcdwealth »

simonw wrote: post #207

As is now traditional for new LLM releases, I used Qwen 3 (32B, run via Ollama on a Mac) to summarize this Hacker News conversation about itself - run at the point when it hit 112 comments. The results were kind of fascinating, because it appeared to confuse my system prompt telling it to summarize the conversation with the various questions asked in the post itself, which it tried to answer. I don't think it did a g…

[dead]

Re: Qwen3: Think deeper, act faster

#267

Post by maybeThrwaway »

sisve wrote: post #248
CSMastermind wrote: post #91

Earlier quoted context omitted.

Right now these models have less censorship than their US counterparts. With that said, they're in a fight for dominance so censoring now would be foolish. If they win and establish a monopoly then the screws will start to turn.

What type of content is removed from US counterparts? Porn, creation of chemical weapons? But not on historical events?

Differ from engine to engine: Googles latest for example put in a few minorities when asking it to create images of nazis. Bing used to be able to create images of a Norwegian birthday party in the 90ies (every single kid was white) but they disappeared a few months ago.

Or you can try to ask them about the grooming scandal in UK. I haven't tried but I have an idea.

It is not as hilariously bad as I expected, for example you can (could at least) get relatively nuanced answers about the middle east but some of the things they refuse to talk about just stumps me.

Re: Qwen3: Think deeper, act faster

#268

Post by redbell »

I'm not sure if it's just me hallucinating, but it seems like with every new model release, it suddenly tops all the benchmark charts—sometimes leaving the competition in the dust. Of course, only real-world testing by actual users across diverse tasks can truly reveal a model’s performance. That said, I still find a sense of excitement and hope for the future of AI every time a new open-source model is released.

Re: Qwen3: Think deeper, act faster

#269

Post by nusl »

redbell wrote: post #268

I'm not sure if it's just me hallucinating , but it seems like with every new model release, it suddenly tops all the benchmark charts—sometimes leaving the competition in the dust. Of course, only real-world testing by actual users across diverse tasks can truly reveal a model’s performance. That said, I still find a sense of excitement and hope for the future of AI every time a new open-source model is released.

Yeah, but their comparison tables appear a bit skewed. o3 doesn't feature, nor does Claude 3.7

Re: Qwen3: Think deeper, act faster

#270

Post by prox »

kaoD wrote: post #256
mrkeen wrote: post #253

Earlier quoted context omitted.

As they say, we shouldn't judge AI by the current state-of-the-art, but by how far and fast it's progressing. I can't wait to see future models get it even more wrong than that.

Personally (anecdata) I haven't experienced any practical progress in my day-to-day tasks for a long time, no matter how good they became at gaming the benchmarks. They keep being impressive at what they're good at (aggregating sources to solve a very well known problem) and terrible at what they're bad at (actually thinking through novel problems or old problems with few sources). E.g. all ChatGPT, Claude and Gemini…

Absolutely, as soon as they hit that mark where things get really specialized, they start failing a lot. They do generalizations on well documented areas pretty good. I only use it for getting a second opinion as it can search through a lot of documents quickly and find me alternative means.
Post reply on HN