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

Qwen3.7-Max: The Agent Frontier

qwen.ai

Post reply on HN
Print view View original

211–220 of 317 posts

Re: Qwen3.7-Max: The Agent Frontier

#211

Post by bel8 »

georgefrowny wrote: post #208
bel8 wrote: post #190

Earlier quoted context omitted.

I use their VSCode extension: https://marketplace.visualstudio.com/items?itemName=sst-dev.... It adds a button to VSCode to open a tab with opencode loaded. It's a bit better than just opening the CLI because it has some vscode integration. With their $10/mo opencode go plan: https://opencode.ai/go For my use it's about endless use of DS4 Flash on high setting. I find high better than max because it's less chatty. Th…

How is that extension compared to, say, DS4 via OpenRouter and the usual VSCode Copilot panel?

Good question.

I haven't tested openrouter but I expect it to be slightly less cheap because it charges per token and opencode Go plan is a $10/mo fixed price model. Economies of scale leads me to think that for heavy use, openrouter will be more costly since opencode Go can subside heavy users like me with money from light users (just like gyms do with people that pay but barely use it).

With that said, I find vscode native copilot chat more pleasant to use, but also more laggy for large sessions.

opencode configuration is less polished and you'll have to grok around for some things. For example opencode CTRL+p conflicts with VSCode CTRL+p. I changed opencode to use Ctrl+L instead.

Re: Qwen3.7-Max: The Agent Frontier

#213

Post by dfansteel »

Mashimo wrote: post #36
dfansteel wrote: post #28

Can anyone check its knowledge base for me? I’m honestly not able to run it and the Qwen models I can run censor information critical towards the Chinese government. Tiananmen Square is the first place to start.

> I’m honestly not able to run it What do you mean? This is not self hosted, it's closed source. And any website that targets China or is hosted in China will probably censor Tiananmen Square.

My computer lacks the ram.

Re: Qwen3.7-Max: The Agent Frontier

#214

Post by ffsm8 »

jampekka wrote: post #193
throawayonthe wrote: post #81

Earlier quoted context omitted.

referencing this: https://artificialanalysis.ai/evaluations/omniscience?models... (had to add it to the chart, wasn't displayed by default. is it the lowest rate in the datasetor no?)

This counts only incorrect answers though. A model can get 0% hallucination rate just by refusing to answer all questions.

Isn't that precisely the reason why we introduced the term hallucination? Because llms have historically always made up bullshit of they cannot answer directly... If they now nailed this to maybe the model not respond instead of responding incorrectly, then a lot of previously unusable usecases would become feasible.

So I feel like that's exactly the right metric and the way to track it wrt hallucinations.

Re: Qwen3.7-Max: The Agent Frontier

#215

Post by speed_spread »

jampekka wrote: post #193
throawayonthe wrote: post #81

Earlier quoted context omitted.

referencing this: https://artificialanalysis.ai/evaluations/omniscience?models... (had to add it to the chart, wasn't displayed by default. is it the lowest rate in the datasetor no?)

This counts only incorrect answers though. A model can get 0% hallucination rate just by refusing to answer all questions.

Yes. A model that can answer "I don't know" would be much more trustable than the current used car salesman we have now.

Re: Qwen3.7-Max: The Agent Frontier

#216

Post by DiabloD3 »

pixelesque wrote: post #150
briga wrote: post #109

I was getting dangerously close to my weekly Claude Code limit last night so I had Claude set up Qwen3.6 with llama.cpp and OpenCode. Honestly it's a great (free!) alternative to Claude Code--certainly more than good enough for a lot of smaller less complex tasks. I'm excited to try this new version. The fact that open-source models are so close to the frontier is very impressive.

Out of interest, what machine and model are you running it on? I tried the qwen3.6-27b Q6_k GUFF in llama.cpp and LM Studio on my M2 MacBook Pro 32GB machine last week, and I barely get a token a second with either. What sort of speed should I be expecting? I tried some of the Llama 3 34b (nous-capybara?) models two years ago with llama.cpp, and I seem to remember getting a few tokens a second then, so not sure if I'…

I recommend sticking with the dense models for both Qwen and Gemma.

On testing I've done on same-quant apples to apples, with F16/F16 (ie, unquantized) kv cache, 35B-A3B underperforms against 27B on anything even remotely complex. But yes, 35B-A3B can be like 3-4x faster on my hardware.

By Qwen's own admission, on any meaningful benchmark (ie, ones that involve logic, math, or tool calling), 27B performs like 122B-10B and 397B-A17B, but 35B-A3B is somewhere between 27B dense and 9B dense.

Also, MTP recently got merged in, so I'd suggest downloading Qwen 3.6 MTP (I assume you get it from unsloth) and updating your copy of llama.cpp, and adding `--spec-type draft-mtp --spec-draft-n-max 2` to your arguments.

https://huggingface.co/unsloth/Qwen3.6-27B-MTP-GGUF/ https://huggingface.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF/

Also, I recommend not quantizing kv cache, and if you do, only quantize v. Lowering model quant while also lowering context size to fit F16/F16 or F16/Q8_0 massively improves model performance for thinking models. Also, quantizing cache, either k or v, decreases speed by a lot on some hardware.

I have a 24gb 7900xtx, so I can fit >32k F16/F16 context with Qwen3.6-27B, but use unsloth's Q3_K_XL. This performs better than Q(4,5,6)_K_XL with v quantized.

Edit: Oh, and since I mentioned Gemma 4, my testing mirrors my Qwen 3.5/3.6 experiences, 26B-A4B performs worse than 31B, but is also way faster. llama.cpp doesn't support Gemma 4's MTP style yet, so both could get even faster.

Re: Qwen3.7-Max: The Agent Frontier

#217

Post by smcleod »

gcr wrote: post #11
smcleod wrote: post #8

Earlier quoted context omitted.

Yeah that 60-150b~ range is such a sweet spot for current 'prosumer' hardware, I'd love to see something like a 120b-a14b or there about.

What’s the price point for getting into that sweet spot? I’m on an M1 Max with 32GB VRAM, so I’m looking forward to the 27B or 35B-A3B models. Is dropping $5k for an RTX 6000 or a DGX Spark really the best option?

Really right now it's the M5 Max MacBook Pro 128GB, the RTX6000 is a nice card but you'd need more than one of them and you have to have a desktop to suit. The DGX Spark is slow and has pretty limited software support.

Re: Qwen3.7-Max: The Agent Frontier

#218

Post by smcleod »

tarruda wrote: post #12
smcleod wrote: post #8

Earlier quoted context omitted.

Yeah that 60-150b~ range is such a sweet spot for current 'prosumer' hardware, I'd love to see something like a 120b-a14b or there about.

I have a 128G mac studio and even 397B was a happy surprise to me due to its high quantization resilience. I've created a 2.54BPW quant that fit on my hardware with 128k context, 20 tps tg and 200tps pp, while maintaining high scores on many benchmarks: https://huggingface.co/tarruda/Qwen3.5-397B-A17B-GGUF/discus...

That's impressive getting a 397B down to <110GB~. HF link is broken though!

Re: Qwen3.7-Max: The Agent Frontier

#219

Post by briga »

wuliwong wrote: post #191
briga wrote: post #109

I was getting dangerously close to my weekly Claude Code limit last night so I had Claude set up Qwen3.6 with llama.cpp and OpenCode. Honestly it's a great (free!) alternative to Claude Code--certainly more than good enough for a lot of smaller less complex tasks. I'm excited to try this new version. The fact that open-source models are so close to the frontier is very impressive.

Do you have a feel for how it Qwen 3.6 compares to Sonnet 4.6? B/C in reality, that's what we use a lot. If we just use Opus 4.7 for everything code related, we'd have a monthly bill 10-20 times higher than using Sonnet where we can.

I would say if Sonnet is a senior engineer, then Qwen3.6 (the 27b model) is probably closer to a junior engineer. Still capable of getting stuff done, just needs more guidance and makes mistakes more often.

Maybe that's underselling it. It is quite a good model and might end up replacing a lot of the work I was sending to Sonnet 4.6.

Also, Sonnet 4.6 is almost certain a much bigger model so the performance differences aren't unexpected.

Re: Qwen3.7-Max: The Agent Frontier

#220

Post by aembleton »

briga wrote: post #109

I was getting dangerously close to my weekly Claude Code limit last night so I had Claude set up Qwen3.6 with llama.cpp and OpenCode. Honestly it's a great (free!) alternative to Claude Code--certainly more than good enough for a lot of smaller less complex tasks. I'm excited to try this new version. The fact that open-source models are so close to the frontier is very impressive.

> Today we introduce Qwen3.7-Max, our latest proprietary model

This is not an open model

Post reply on HN