# Qwen 3.8 is about to launch, open weight too **URL:** \ **Category:** DGX Spark / GB10\ **Created:** [July 19, 2026, 12:07pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396 "2026-07-19T12:07:42Z")\ **Posts on this page:** 20\ **Page:** 1
**Author:** ![0rand](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/0rand/32/512004_2.png) [@0rand](https://forums.developer.nvidia.com/u/0rand)\ **Post date:** [July 19, 2026, 12:07pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/1 "2026-07-19T12:07:42Z")
![Screenshot_20260719_140433_Firefox](https://global.discourse-cdn.com/nvidia/original/4X/8/d/5/8d5518affff61975f413ab8aed8108d03a7df01c.jpeg) 2+ trillion, about to open weight. Maybe a small companion but not very likely imo. ![Screenshot_20260719_140719_Firefox](https://global.discourse-cdn.com/nvidia/original/4X/9/8/b/98b39dd794772eb2225d6eaecc3da8d62edeb513.jpeg) ---
**Author:** ![renek](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@renek](https://forums.developer.nvidia.com/u/renek)\ **Post date:** [July 19, 2026, 12:33pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/2 "2026-07-19T12:33:12Z")
i talked to reps in Europe and they dont want to release the 122B+ models that would fit perfectly for us… of course might be 35B MoE and 27B Dense, these would still be candidates to challenge DS4 Flash GA, Mimo v2.5 and Minimax M3. Looking forward to it and maybe they will surprise us… ---
**Author:** ![jbourny](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@jbourny](https://forums.developer.nvidia.com/u/jbourny)\ **Post date:** [July 19, 2026, 5:43pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/3 "2026-07-19T17:43:08Z")
They will surprise us by putting on Huggingface only the model of 2T that no one will be able to load, I think it’s over the little models it doesn’t help them to do it… ---
**Author:** ![azampatti](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/azampatti/32/503643_2.png) [@azampatti](https://forums.developer.nvidia.com/u/azampatti)\ **Post date:** [July 20, 2026, 2:13pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/4 "2026-07-20T14:13:44Z")
I need a new 122b sooo bad :) Like I said before, a mid-point would be great too, like 80B-A8B or something along those lines, that might be even better :) ---
**Author:** ![Teason2026](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@Teason2026](https://forums.developer.nvidia.com/u/Teason2026)\ **Post date:** [July 20, 2026, 2:39pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/5 "2026-07-20T14:39:24Z")
small models these days coming from google, but optimized for phones :) frontier now 1T+ parameters, so, why would they invest time to build small, if they can build big and someone will distill it to old horses. ---
**Author:** ![Digital\_David](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/digital_david/32/495104_2.png) [@Digital\_David](https://forums.developer.nvidia.com/u/Digital_David)\ **Post date:** [August 4, 2026, 3:21pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/6 "2026-08-04T15:21:55Z")
> New 27B dense is on the way, holding out hopes for a 35B-A3b MOE… ---
**Author:** ![2893f57acbe7739fc460e9bd7](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@2893f57acbe7739fc460e9bd7](https://forums.developer.nvidia.com/u/2893f57acbe7739fc460e9bd7)\ **Post date:** [August 4, 2026, 5:04pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/7 "2026-08-04T17:04:31Z")
Agreed, seriously holding out hope for a 122B on 3.8, but some other mid quants would be nice too. I just seriously hope it is not just “benchmaxed”. ---
**Author:** ![secuspec](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/secuspec/32/526085_2.png) [@secuspec](https://forums.developer.nvidia.com/u/secuspec)\ **Post date:** [August 8, 2026, 5:21pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/8 "2026-08-08T17:21:02Z")
Currently, at least from my perspective, the 122B Parameter Model serves me perfectly. It really seems the sweet spot on a singe DGX. Perhaps we’ll get a 3.8 successor - who knows? ---
**Author:** ![renek](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@renek](https://forums.developer.nvidia.com/u/renek)\ **Post date:** [August 8, 2026, 5:26pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/9 "2026-08-08T17:26:00Z")
for now what ive seen only 27B dense is confirmed ---
**Author:** ![0rand](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/0rand/32/512004_2.png) [@0rand](https://forums.developer.nvidia.com/u/0rand)\ **Post date:** [August 8, 2026, 6:12pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/10 "2026-08-08T18:12:41Z")
I think they correctly address current ram/dram shortage and target most accessible size that performs well: 27b in int4 can fit 24gb cards and mid-tier apple (32-64gb), AMD lpdramm platforms ---
**Author:** ![2893f57acbe7739fc460e9bd7](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@2893f57acbe7739fc460e9bd7](https://forums.developer.nvidia.com/u/2893f57acbe7739fc460e9bd7)\ **Post date:** [August 8, 2026, 6:17pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/11 "2026-08-08T18:17:19Z")
If 27B will be a huge step up in how well it performs vs its 3.6 version, there might not be a point in my Spark anymore, to be honest. My 5090 runs 27B with 200k context and pretty good performance on just a plain jane llama.cpp installation. I’m sure a vllm rendition would perform even better. Currently 3.5 122B is still the best model I have found in general intelligence for my task. Laguna I’m mystified about because it’s like it performs incredibly well… once. Then never again. 35 A3B is okay but kind of pointless when 122B exists (as far as quality), and my experience with lower quant 731 versions of DSV4F have been mixed. Maybe the 731 version has been minmaxed for coding tasks. Not sure what to do on a single spark anymore. That’s why I hope they release a 122B or some variant in between, otherwise it’s going to feel kind of unnecessary soon. ---
**Author:** ![0rand](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/0rand/32/512004_2.png) [@0rand](https://forums.developer.nvidia.com/u/0rand)\ **Post date:** [August 8, 2026, 6:22pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/12 "2026-08-08T18:22:05Z")
Best reason to own one spark is that it’s a half of 2 node cluster. That’s a game changer. I have m5 max 128gb but mostly use 27b and 35b there - better than 122b if you can plug a cloud model for tough questions, leaves more run for cache and many other apps. BTW checked today gigabyte atop top 4T price, same as I own, same seller, +60% from what I paid for my second spark 3 months ago ---
**Author:** ![2893f57acbe7739fc460e9bd7](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@2893f57acbe7739fc460e9bd7](https://forums.developer.nvidia.com/u/2893f57acbe7739fc460e9bd7)\ **Post date:** [August 9, 2026, 6:55am UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/13 "2026-08-09T06:55:30Z")
Well at that point, I’m going to be around 7k deep into the platform to run however much better DSV4F is. My current spark cost me 3k and another one will cost 4k. I guess we’ll have to see. I’m still holding out hope for a 122B quant (or something between 35 and 122). ---
**Author:** ![ark4](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@ark4](https://forums.developer.nvidia.com/u/ark4)\ **Post date:** [August 12, 2026, 8:43am UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/14 "2026-08-12T08:43:33Z")
I’d love a 122b MoE to run on 2x DGX sparks as a cluster. Hoping they bring one out. ---
**Author:** ![mashie](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/mashie/32/503632_2.png) [@mashie](https://forums.developer.nvidia.com/u/mashie)\ **Post date:** [August 12, 2026, 9:12am UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/15 "2026-08-12T09:12:38Z")
I rather have a 400-500B MoE version for 4-node spark clusters. The 2-node clusters are already covered with DSv4 Flash. ---
**Author:** ![ark4](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@ark4](https://forums.developer.nvidia.com/u/ark4)\ **Post date:** [August 12, 2026, 9:47am UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/16 "2026-08-12T09:47:39Z")
That would be good too, but i don’t have a CX7 switch yet to support 4 DGX sparks :( ---
**Author:** ![renek](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@renek](https://forums.developer.nvidia.com/u/renek)\ **Post date:** [August 12, 2026, 9:58am UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/17 "2026-08-12T09:58:39Z")
i hope they go for all size, would be good even to have new 9B and 4B Versions ---
**Author:** ![voktolom](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/voktolom/32/504559_2.png) [@voktolom](https://forums.developer.nvidia.com/u/voktolom)\ **Post date:** [August 12, 2026, 11:34am UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/18 "2026-08-12T11:34:52Z")
> **[ModelScope 魔搭社区](https://modelscope.cn/models/Qwen/Qwen3.8-2.4T-A95B?spm=a2ty_o01.29997173.0.0.191d55fbcxiZyQ&file=Qwen3.8-2.4T-A95B)** > > ModelScope——汇聚各领域先进的机器学习模型，提供模型探索体验、推理、训练、部署和应用的一站式服务。在这里，共建模型开源社区，发现、学习、定制和分享心仪的模型。 There are three hours left to wait) And once again, sleepless nights)) Along with the big one, qwen3.8-27B should be released as well. ---
**Author:** ![jbourny](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@jbourny](https://forums.developer.nvidia.com/u/jbourny)\ **Post date:** [August 12, 2026, 12:43pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/19 "2026-08-12T12:43:32Z")
After the launch it will not be forgotten that it will take several more days to actually use it on Dgx Spark because the BF16 will run at only 5/10 tok/s. We’ll have to wait few days for NVFP4 quantization and Dflash to have at least 20tok/s and several more days to have a functional JINJA template because as always Qwen releases a model with defective templates. I don’t think we can just use the template of 3.6 because in 3.8 there seems to be the reasoning effort system, there must be specific switches in the templates I think. ---
**Author:** ![jwarner](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@jwarner](https://forums.developer.nvidia.com/u/jwarner)\ **Post date:** [August 12, 2026, 12:53pm UTC](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396/20 "2026-08-12T12:53:43Z")
Make your own quant in a few hours. It won’t take days. [Next page](https://forums.developer.nvidia.com/t/qwen-3-8-is-about-to-launch-open-weight-too/377396.md?page=2)