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

Qwen 3.8 27B

huggingface.co

Post reply on HN
Print view View original

471–480 of 848 posts

Re: Qwen 3.8 27B

#471

Post by tinfoilhatter »

williamcotton wrote: post #247
tinfoilhatter wrote: post #196

Earlier quoted context omitted.

So if I hire an artist and am a motivated individual, have an idea for a painting, have background knowledge about paintings and have taste in paintings and can offer a critique of the painting as the artist paints it, then somehow I created the painting? Absurd logic. The AI built the website.

There have been plenty of workshops where artists hire assistant painters while maintaining authorship over the works themselves, from Rembrandt to Warhol to Hirst.

Sure - and a factual statement would be, these artists hired assistant artists to help them create their paintings. Just like it would be factual to say that the person directed a LLM to build an app for them.

Re: Qwen 3.8 27B

#472

Post by orangedog »

simplyluke wrote: post #84
ramon156 wrote: post #64

People will claim it's not comparable to Opus despite it beating the score. I'm not sure I disagree, but I'm also unsure whether I care. Most new models nowadays are "good enough". I cannot complain because I'd rather spend that time improving my prompts and docs. Opus might be a _slight bit better_ at picking up vague hints, but it's also extremely expensive, and I hit the 5 hour limit way too quick. I care a lot ab…

I'm convinced a lot of the anti-open-weight model comments at this point are inorganic traffic - there's trillions in investor money riding on a world where these models aren't cheap commodities. Having actually used things like the recent GLM, Kimi, and Qwen I think any edge the labs have is marginal at most and actually prefer the open weight models in most day to day usage. Anthropic's recent releases are wordy to…

Why are you convinced of that? Pretty much every time I see statements like that online, I can find plenty of organic traffic supporting it, not everything is a bot.

I just wouldn't bias myself that way, most people haven't really used local models. This stuff is pretty much all subjective evaluation, there's plenty of reasons for people to favor certain models or disfavor others.

Re: Qwen 3.8 27B

#473

Post by MrDrMcCoy »

Der_Einzige wrote: post #180
skrebbel wrote: post #178

Earlier quoted context omitted.

I'd love to understand this more. Are you saying the Qwen team spends their very impressive human and compute resources on publishing these amazing models and then botches the chat template with mundane bugs? Like maybe I just misunderstand what's the hard part but wouldn't you assume that people who can put together an impressive model can also write a proper jinja chat template for it?

Yes yes, oh god yes. They also spread FUD in the form of terrible recommended sampler settings. If you're using llamacpp, turn on top-n-sigma with sigma of 1, turn off top-p/top-k. You'll thank me later.

For those us us who don't know, what do those parameters do and why are they better?

Re: Qwen 3.8 27B

#474

Post by mirekrusin »

redox99 wrote: post #143
NitpickLawyer wrote: post #34

Earlier quoted context omitted.

> Beats Opus 4.7 Max I'm a huge open model fan, and have used them since forever, even have daily drivers for on-prem dev, but no. They do not beat opus on real-world usage. Qwen models are impressively good for what they are, are "good enough" for plenty tasks, can be ran locally on decently priced hardware, and so on. They certainly have their uses, and the field in general has advanced faster than my early expecta…

Yep. These small models are actually worse than GPT 3.5 at some tasks (like recalling facts). You can definitely make models smarter at specific tasks (like tool calling, coding) but you can't compress the entire human knowledge into a 30GB file. It's just not enough bits.

That's a good sign.

Ideal local model would not know stuff like who Britney Spears is, best to leave precious weights for something useful.

Of course the line is very blurry but I'd be perfectly happy with local model that doesn't know anything about history, geopolitics, art or even biology etc. just coding, operating systems etc.

Re: Qwen 3.8 27B

#475

Post by lacoolj »

simonw wrote: post #419

Absolutely the best pelican I've seen from a model that runs on my laptop: https://tools.simonwillison.net/markdown-svg-renderer#url=ht... Bicycle is the right shape. Pelican beak is excellent. Nice background. Most importantly, the pelican has one leg on each side of the bicycle - that's very rare. (No chain on this bicycle though - in the reasoning trace it says "already chainstay... skip chain detail; maybe a smal…

Totally agree. I actually didn't know you already got to it and I ran it on mine. It even added a scarf lol

https://imgur.com/a/IgYKnmk

Took almost 90 minutes on my Threadripper Pro, and feels well worth it.

28k tokens from the Q8 model

Re: Qwen 3.8 27B

#476

Post by seanmcdirmid »

walrus01 wrote: post #470
seanmcdirmid wrote: post #462

Earlier quoted context omitted.

My use cases try to avoid accessing world knowledge in the model (I give it access to web search for some adhoc RAG), and ya, I'm just focused on coding so that's the only place I'm looking at right now.

I think you may find that the dense 27B also does better if challenged with more rare coding tasks, less common or weird languages or things that aren't well represented in the active 3B parameters of the MoE model (eg: NOT css, javascript, python, c++, etc).

True. Benchmarks have led me astray before, I'll have to actually benchmark it directly in Goose with my real use cases right now (which are mostly writing python code, so it probably still doesn't apply?). Still, I've never noticed a benefit to using a dense model rather than an MoE for coding with an agent tasks.

Re: Qwen 3.8 27B

#477

Post by gpt5 »

EB66 wrote: post #448
stusmall wrote: post #436

Earlier quoted context omitted.

I'm generally extremely skeptical about a lot of the model hype that show up in comments. Except when there is an extreme mismatch the performance, quirks and quality of these things are difficult to nail down. You wouldn't know that from the comment section of every single release . I think some of these are excited, eager users always ready to hype up the new thing. The same crowd that previously would constantly p…

> I think most of it is bot driven spam by the various labs. It's hard not to notice 3 month old accounts with very strong opinions about various frontier labs and little else. I've assumed the same as well. I also assume that many of the companies developing these models engage in benchmaxxing. At my company we've developed our own internal benchmarks for evaluating LLM models as they become available. The benchmark…

It actually is showing in public benchmark if you know how to look for it. For example, in Terminal-Bench 2.1, GLM 5.2 received 78%, while GPT 5.6 Sol received 88%.

Then Terminal-Bench 3.0 came out (where the questions are new), and GPT 5.6 Sol received 34.6%, while GLM 5.2 dropped to a whopping 4.6%.

Re: Qwen 3.8 27B

#478

Post by reilly3000 »

fragmede wrote: post #426
bilekas wrote: post #380

Earlier quoted context omitted.

I've worked with a good few types of managers of the years, none would take ownership of my work. The end result? Sure, for shareholders/this managers that is fine. But that's not the same as me being needed to be on a call with integration teams and having to run the call. It's just a layer of abstraction for management. I own the creation side of the product. Who created it is still me.

Did Steve Jobs create the iPhone?

Would it exist in it’s present form if he didn’t exist?

Re: Qwen 3.8 27B

#479

Post by bilekas »

MrDrMcCoy wrote: post #458
bilekas wrote: post #405

Earlier quoted context omitted.

I'll actually defend you on this one, but not only the command, that's more of a fine tuning option. On the hugging face link for example, there's no flat 'download' area. It's completely unintuitive for people to know how to use it. Reminds me of my first time trying to make sense of torrents when they first arrived. There are some clients that will index the models and allow you to do that but I'm no expert, I've u…

Official llama.cpp releases ship with huggingface support. If you don't want to download it yourself, you can just use the `repo/model:quant` convention and it will handle downloading locally for you.

But you're assuming I'm using the Olama studio. This model as far as I see doesn't have a gguf download.. Unless I'm missing something on the page.

If I want to download the model myself, it's not clear. I thought it was supposed to behave like a package manager. But even in nuGet I can download a zip of the package.

Re: Qwen 3.8 27B

#480

Post by c7b »

For those commenting on the long reasoning, it may be interesting to know that the reasoning effort is set to xhigh by default [0]. Other possible values are medium, low and none. Flag for changing it in llama.cpp below, but note that the long reasoning seems to contribute a great deal to the quality.

  --chat-template-kwargs '{"preserve_thinking":true,"reasoning_effort":"medium"}'
[0] https://unsloth.ai/docs/models/qwen3.8#thinking--preserve-th...
Post reply on HN