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

Qwen3.6-35B-A3B: Agentic coding power, now open to all

qwen.ai

Post reply on HN
Print view View original

451–460 of 563 posts

Re: Qwen3.6-35B-A3B: Agentic coding power, now open to all

#451

Post by naasking »

Der_Einzige wrote: post #172
ekianjo wrote: post #124

Earlier quoted context omitted.

yeah the 27B feels like something completely different. If you use it on long context tasks it performs WAY better than 35b-a3b

I've been telling analysts/investors for a long time that dense architectures aren't "worse" than sparse MoEs and to continue to anticipate the see-saw of releases on those two sub-architectures. Glad to continuously be vindicated on this one. For those who don't believe me. Go take a look at the logprobs of a MoE model and a dense model and let me know if you can notice anything. Researchers sure did.

MoE isn't inherently better, but I do think it's still an under explored space. When your sparse model can do 5 runs on the same prompt in the same time as a dense model takes to generate one, there opens up all sorts of interesting possibilities.

Re: Qwen3.6-35B-A3B: Agentic coding power, now open to all

#452

Post by naasking »

3836293648 wrote: post #374

Qwen3.6 and Gemma4 have the same issue of never getting to the point and just getting stuck in never ending repeating thought loops. Qwen3.5 is still the best local model that works.

Quantization can introduce these issues, and Gemma 4 also had issues because the prompt tokens that Gemma used was new and not well supported yet.

Re: Qwen3.6-35B-A3B: Agentic coding power, now open to all

#453

Post by naasking »

coder543 wrote: post #183
wild_egg wrote: post #146

Earlier quoted context omitted.

Where did you see a haiku comparison? Haiku 4.5 was my daily driver for a month or so before Opus 4.5 dropped and would be unreasonably happy if a local model can give me similar capability

Artificial Analysis hasn't posted their independent analysis of Qwen3.6 35B A3B yet, but Alibaba's benchmarks paint it as being on par with Qwen3.5 27B (or better in some cases). Even Qwen3.5 35B A3B benchmarks roughly on par with Haiku 4.5, so Qwen3.6 should be a noticeable step up. https://artificialanalysis.ai/models?models=gpt-oss-120b%2Cg... No, these benchmarks are not perfect, but short of trying it yourself,…

Qwen models commonly get accused of benchmaxxing though. Just something to keep in mind when weighing the standard benchmarks.

Re: Qwen3.6-35B-A3B: Agentic coding power, now open to all

#454

Post by mistercheese »

aliljet wrote: post #73

I'm broadly curious how people are using these local models. Literally, how are they attaching harnesses to this and finding more value than just renting tokens from Anthropic of OpenAI?

I use local models for asking about personal financial or health data that I want to keep local and private. Or even just whipping up quick and dirty prototypes for whatever I can think of but not seriously enough to spend tokens that I rather use on real projects.

Re: Qwen3.6-35B-A3B: Agentic coding power, now open to all

#455

Post by mistercheese »

seemaze wrote: post #417
3836293648 wrote: post #377

Earlier quoted context omitted.

How much VRAM do you need for that?

I squeeze Qwen3.5-122B-A10B at Q6 into 128GB. It's a great model.

Wow what kind of hardware do you have? Mac Studio, dgx spark, strix halo? How fast is it?

Re: Qwen3.6-35B-A3B: Agentic coding power, now open to all

#456

Post by mistercheese »

zozbot234 wrote: post #215
deaux wrote: post #211

Earlier quoted context omitted.

While they can be run locally, and most of the discussion on HN about that, I bet that if you look at total tok/day local usage is a tiny amount compared to total cloud inference even for these models. Most people who do use them locally just do a prompt every now and then.

This is why I'd like to see a lot more focus on batched inference with lower-end hardware. If you just do a tiny amount of tok/day and can wait for the answer to be computed overnight or so, you don't really need top-of-the-line hardware even for SOTA results.

That’s a good point. I think I saw Together.ai with that offering, but for some reason just never think to throw random non urgent coding tasks at it overnight

Re: Qwen3.6-35B-A3B: Agentic coding power, now open to all

#457

Post by npodbielski »

I am not sure. I tested it locally on my Desktop Framework and it so far it seem to giving me worse answers then Qwen 3.5. Maybe it is because I am chatting with models in my language instead of enlish or maybe it is optimised for coding instead.

I asked it to give me instruction on how to create SSH key and it tried to do it instead of just answering.

https://internetexception.com/2026/04/16/testing-qwen-3-6/

Re: Qwen3.6-35B-A3B: Agentic coding power, now open to all

#458

Post by jmb99 »

kylehotchkiss wrote: post #252
zozbot234 wrote: post #40

Earlier quoted context omitted.

This is just one model in the Qwen 3.6 series. They will most likely release the other small sizes (not much sense in keeping them proprietary) and perhaps their 122A10B size also, but the flagship 397A17B size seems to have been excluded.

How many people/hackernews can run a 397b param model at home? Probably like 20-30.

I’ve mentioned this as an option in other discussions, but if you don’t care that much about tok/sec, 4x Xeon E7-8890 v4s with 1TB of DDR3 in a supermicro X10QBi will run a 397b model for Full (non-quantized, non-distilled) DeepSeek runs at 1-2 tok/sec. A model half the size would probably be a little faster. This is also only with the basic NUMA functionality that was in llama.cpp a few months ago, I know they’ve added more interesting distribution mechanisms recently that I haven’t had a chance to test yet.

Re: Qwen3.6-35B-A3B: Agentic coding power, now open to all

#459

Post by blurbleblurble »

kylehotchkiss wrote: post #252
zozbot234 wrote: post #40

Earlier quoted context omitted.

This is just one model in the Qwen 3.6 series. They will most likely release the other small sizes (not much sense in keeping them proprietary) and perhaps their 122A10B size also, but the flagship 397A17B size seems to have been excluded.

How many people/hackernews can run a 397b param model at home? Probably like 20-30.

It only has 17b active params, it's a mixture of experts model. So probably a lot more people than you realize!

Re: Qwen3.6-35B-A3B: Agentic coding power, now open to all

#460

Post by lpnam0201 »

kridsdale3 wrote: post #272
kylehotchkiss wrote: post #252

Earlier quoted context omitted.

How many people/hackernews can run a 397b param model at home? Probably like 20-30.

I can (barely, but sustainably) run Q3.5 397B on my Mac Studio with 256GB unified. It cost $10,000 but that's well within reach for most people who are here, I expect.

In where I am living, 10k USD is a little more than 3 years worth of rent, for a relatively new and convenient 2 bedroom apartment.
Post reply on HN