Загрузка видео...

Не удалось загрузить видео

Возникла проблема при загрузке этого видео. Это может быть связано с временными проблемами сети или видео может быть недоступно.

На главную

SemiAnalysis tested Qwen3.8-Max-Preview 2.4T params

SemiAnalysis

176,331 subscribers

58,140 просмотров • 2 месяцев назад •via X (Twitter)

Наука и технологии

Anya Rossi• Live Now

Private livecam show

Комментарии: 7

Фото профиля Coyote

Coyote2 месяцев назад

Wait. The model is officially out on August 2 or your own tests of the model from the web app?

Фото профиля SemiAnalysis

SemiAnalysis2 месяцев назад

We got early access

Фото профиля volar yu

volar yu2 месяцев назад

ಠ_ಠ I thought they were just benchmaxxing the prev one this one seems better?

Фото профиля rakia ahmad

rakia ahmad2 месяцев назад

ok can you test the new qwen 3.8 max not preview ?

Фото профиля gus mann

gus mann2 месяцев назад

Canceled video immediately, welcome relief from the soul-sucking fake "music".

Фото профиля ZHUOLIN

ZHUOLIN2 месяцев назад

A full preview test will be useful once the open-weight version is available.

Фото профиля Bhaskar Deol

Bhaskar Deol2 месяцев назад

Worth revisiting when the weights land next week. Today it is one vendor at one price, so there is nothing to compare. On GLM 5.2 right now there are thirty live endpoints, and the dearest input price is a little over three times the cheapest.

Похожие видео

Kolejna grupa modeli po testach ... Open Source na łopatkach totalnie .... - gpt-5.6 sol - jako baseline - 4.42$ - nemotron 3.5 lightning - 0.04$ - deepseek-v4-pro-0813 - 12.20$ - Qwen3.8 Max - 11.72$ - Qwen3.8 2.4T A95B - 14.81$

Kolejna grupa modeli po testach ... Open Source na łopatkach totalnie .... - gpt-5.6 sol - jako baseline - 4.42$ - nemotron 3.5 lightning - 0.04$ - deepseek-v4-pro-0813 - 12.20$ - Qwen3.8 Max - 11.72$ - Qwen3.8 2.4T A95B - 14.81$

Remek Kinas

30,103 просмотров • 2 месяцев назад

Alibaba made Qwen 3.8 Max free on Qoder Desktop for 14 days it's their 2.4T parameter model. rivals Fable 5 on coding benchmarks. runs in a desktop IDE. what you get for $0: -2.4 trillion parameters -Qwen3.8-Max-Preview in a full agentic IDE -code completion, test generation, file editing -2-week Pro trial, then 90-98% off preview pricing what this replaces: -Fable 5 at $30/M output tokens -$20/mo Pro coding subs how to set it up (2 min): 1. Download at 2. Sign up with email 3. Add Qwen3.8-Max-Preview in Settings → Models Works on macOS, Windows, Linux. Important: 2-week free trial. No card needed. Preview pricing after is 90-98% off. Use it before they take it off

Alibaba made Qwen 3.8 Max free on Qoder Desktop for 14 days it's their 2.4T parameter model. rivals Fable 5 on coding benchmarks. runs in a desktop IDE. what you get for $0: -2.4 trillion parameters -Qwen3.8-Max-Preview in a full agentic IDE -code completion, test generation, file editing -2-week Pro trial, then 90-98% off preview pricing what this replaces: -Fable 5 at $30/M output tokens -$20/mo Pro coding subs how to set it up (2 min): 1. Download at 2. Sign up with email 3. Add Qwen3.8-Max-Preview in Settings → Models Works on macOS, Windows, Linux. Important: 2-week free trial. No card needed. Preview pricing after is 90-98% off. Use it before they take it off

painn

112,969 просмотров • 3 месяцев назад

Just tested Qwen3.8-Max-Preview on my benchmark. It got absolutely cooked. Not Fable 5 level. Not even remotely close. It doesn't come close to GPT-5.6 or Opus 4.8 Max either. Honestly, marketing it at that level feels incredibly misleading. I wasted my time and money testing this. Qwen, refund me.

Just tested Qwen3.8-Max-Preview on my benchmark. It got absolutely cooked. Not Fable 5 level. Not even remotely close. It doesn't come close to GPT-5.6 or Opus 4.8 Max either. Honestly, marketing it at that level feels incredibly misleading. I wasted my time and money testing this. Qwen, refund me.

OmedTheVibeCoder

184,120 просмотров • 3 месяцев назад

First impression of how fast the new Qwen3.8-Flash-Next runs on a single DGX Spark. 180B params, 262k context, all local.

First impression of how fast the new Qwen3.8-Flash-Next runs on a single DGX Spark. 180B params, 262k context, all local.

0xBakeer

22,118 просмотров • 2 месяцев назад

Meet Qwen3.8-Max: A New Bar for Coding and Cowork.

Meet Qwen3.8-Max: A New Bar for Coding and Cowork.

Qwen

7,708,257 просмотров • 2 месяцев назад

Qwen3.8 max NYC skate game in a single C++ file

Qwen3.8 max NYC skate game in a single C++ file

BijanBowen

28,168 просмотров • 2 месяцев назад

What happens when AI codes autonomously for 10+ days? 🚀 Qwen3.8‑Max built, tested, verified, and repaired a self‑evolving harness, all in one long‑horizon autonomous coding run. Watch it build from scratch!

What happens when AI codes autonomously for 10+ days? 🚀 Qwen3.8‑Max built, tested, verified, and repaired a self‑evolving harness, all in one long‑horizon autonomous coding run. Watch it build from scratch!

Alibaba Group

2,206,138 просмотров • 2 месяцев назад

We tested Qwen3.8-Flash-Next on the MINISFORUM MS-S1 MAX-P495 AI Workstation, powered by AMD Ryzen™ AI Max+ PRO 495. 🤖 🚀 🛒 Learn more: 🇺🇸 US: 🇪🇺 EU: What would you like to test on it? #MINISFORUM #MSS1MAXP495 #LocalAI #AIWorkstation

We tested Qwen3.8-Flash-Next on the MINISFORUM MS-S1 MAX-P495 AI Workstation, powered by AMD Ryzen™ AI Max+ PRO 495. 🤖 🚀 🛒 Learn more: 🇺🇸 US: 🇪🇺 EU: What would you like to test on it? #MINISFORUM #MSS1MAXP495 #LocalAI #AIWorkstation

MINISFORUM Official

13,180 просмотров • 25 дней назад

Qwen3.8-Max is built for more than just answering prompts. Alibaba’s newest flagship model brings: • 2.4T total parameters • 1M-token context • Native vision + text reasoning • 95B active parameters through Sparse MoE The interesting part is what it can actually do. From autonomous coding and long-running tasks to analyzing 200+ page documents and massive streams of information, Qwen3.8-Max is designed to work through complex tasks, not just generate text. Now accessible globally through Alibaba Cloud Model Studio. Alibaba Group Qwen Qwen Studio: API: Blog: Alizila:

Qwen3.8-Max is built for more than just answering prompts. Alibaba’s newest flagship model brings: • 2.4T total parameters • 1M-token context • Native vision + text reasoning • 95B active parameters through Sparse MoE The interesting part is what it can actually do. From autonomous coding and long-running tasks to analyzing 200+ page documents and massive streams of information, Qwen3.8-Max is designed to work through complex tasks, not just generate text. Now accessible globally through Alibaba Cloud Model Studio. Alibaba Group Qwen Qwen Studio: API: Blog: Alizila:

Samaira AI

81,746 просмотров • 2 месяцев назад

Before Qwen3.8-Max-0902 landed in Arena yesterday, we met Qwen3.8-27B earlier this month. While in SF last week, Peter Gostev shared his technical assessment of this smaller open-source model from Qwen. Find the full video covering Qwen3.8-27B below.

Before Qwen3.8-Max-0902 landed in Arena yesterday, we met Qwen3.8-27B earlier this month. While in SF last week, Peter Gostev shared his technical assessment of this smaller open-source model from Qwen. Find the full video covering Qwen3.8-27B below.

Arena.ai

27,788 просмотров • 1 месяц назад

📢 Qwen's Qwen3.8-2.4T-A95B weights just dropped, and SGLang already has day-0 support with DSpark. HF: Cookbook: 🖖 What kind of model can help optimize an inference engine? Qwen3.8-2.4T-A95B can take on complex engineering tasks, plan and iterate on its own, and keep going until the job is done. We even asked it to optimize its own serving on SGLang, and it ran unattended for 4.5 hours with verified performance improvements. Thanks to the Qwen team Qwen, NVIDIA AI, and AI at AMD for the close collaboration.

📢 Qwen's Qwen3.8-2.4T-A95B weights just dropped, and SGLang already has day-0 support with DSpark. HF: Cookbook: 🖖 What kind of model can help optimize an inference engine? Qwen3.8-2.4T-A95B can take on complex engineering tasks, plan and iterate on its own, and keep going until the job is done. We even asked it to optimize its own serving on SGLang, and it ran unattended for 4.5 hours with verified performance improvements. Thanks to the Qwen team Qwen, NVIDIA AI, and AI at AMD for the close collaboration.

SGLang

27,317 просмотров • 2 месяцев назад

We ran a test between the new Qwen3.8-Max, Opus 5 and GPT-5.6 Sol. 3 models. same prompt. one-shot with the /design command. Reviewed gameplay features, UX/UI and cost. 🔹 Qwen3.8-Max → 9/10 · $0.0248 🔹 GPT-5.6 Sol → 9/10 · $0.150 🔹 Opus 5 → 8.5/10 · $0.253 Qwen3.8-Max is approximately 4.2× cheaper than GPT-5.6, Sol, Opus 5 and has the same level of UI, UX, and gameplay.

We ran a test between the new Qwen3.8-Max, Opus 5 and GPT-5.6 Sol. 3 models. same prompt. one-shot with the /design command. Reviewed gameplay features, UX/UI and cost. 🔹 Qwen3.8-Max → 9/10 · $0.0248 🔹 GPT-5.6 Sol → 9/10 · $0.150 🔹 Opus 5 → 8.5/10 · $0.253 Qwen3.8-Max is approximately 4.2× cheaper than GPT-5.6, Sol, Opus 5 and has the same level of UI, UX, and gameplay.

Command Code

180,367 просмотров • 2 месяцев назад

i tested Qwen3.8 27B locally on my M4 Pro to see how it behaves as a coding agent inside my project showing step-by-step the full guide on how you can run it, perfect for your first local model 00:00 Why I’m testing Qwen3.8 00:49 What Qwen3.8 includes 00:59 Qwen3.8 Max vs 27B 01:14 Qwen3.8 Max explained 02:13 Qwen’s benchmark claims 02:31 Its autonomous coding project 03:18 Qwen3.8 27B explained 04:30 Why I’m testing it locally 05:05 Download Qwen 3.8 in LM Studio 06:00 Pick the Q4 model 06:41 Run it as a local server 07:33 How fast is it? 08:36 Set up OpenCode 09:33 Connect LM Studio to OpenCode 10:03 Add Qwen 3.8 to OpenCode 11:34 Match the context settings 15:00 Give Qwen a real coding task 15:49 A mistake I caught live: session compaction 18:00 Review the code 19:30 My verdict

i tested Qwen3.8 27B locally on my M4 Pro to see how it behaves as a coding agent inside my project showing step-by-step the full guide on how you can run it, perfect for your first local model 00:00 Why I’m testing Qwen3.8 00:49 What Qwen3.8 includes 00:59 Qwen3.8 Max vs 27B 01:14 Qwen3.8 Max explained 02:13 Qwen’s benchmark claims 02:31 Its autonomous coding project 03:18 Qwen3.8 27B explained 04:30 Why I’m testing it locally 05:05 Download Qwen 3.8 in LM Studio 06:00 Pick the Q4 model 06:41 Run it as a local server 07:33 How fast is it? 08:36 Set up OpenCode 09:33 Connect LM Studio to OpenCode 10:03 Add Qwen 3.8 to OpenCode 11:34 Match the context settings 15:00 Give Qwen a real coding task 15:49 A mistake I caught live: session compaction 18:00 Review the code 19:30 My verdict

Nidhi Singh

43,169 просмотров • 2 месяцев назад

Alibaba is casually letting everyone try its newest flagship model for free 😳 Qwen3.8-Max-Preview is live right now on no API key no credit card just open the site, select the model, and start using it what you get for $0: - Alibaba's latest flagship reasoning model - 2.4T parameter MoE - built-in Web Dev mode - generates complete websites from a single prompt - free access directly in your browser why this is worth checking: > build and preview websites instantly > test Alibaba's latest frontier model without paying > compare it with your current coding workflow before committing to an API getting started takes less than a minute: 1. go to 2. sign in with a free account 3. select "Qwen3.8-Max-Preview" 4. enable Web Dev mode 5. describe what you want to build I literally typed "hello" and it generated an interactive landing page with animations, responsive design, and live preview important: - currently in preview - Alibaba hasn't released official benchmark results yet - treat any performance claims you see online as unverified until official numbers are published A simple way to try Alibaba's latest flagship AI model for free before API access becomes the main way to use it

Alibaba is casually letting everyone try its newest flagship model for free 😳 Qwen3.8-Max-Preview is live right now on no API key no credit card just open the site, select the model, and start using it what you get for $0: - Alibaba's latest flagship reasoning model - 2.4T parameter MoE - built-in Web Dev mode - generates complete websites from a single prompt - free access directly in your browser why this is worth checking: > build and preview websites instantly > test Alibaba's latest frontier model without paying > compare it with your current coding workflow before committing to an API getting started takes less than a minute: 1. go to 2. sign in with a free account 3. select "Qwen3.8-Max-Preview" 4. enable Web Dev mode 5. describe what you want to build I literally typed "hello" and it generated an interactive landing page with animations, responsive design, and live preview important: - currently in preview - Alibaba hasn't released official benchmark results yet - treat any performance claims you see online as unverified until official numbers are published A simple way to try Alibaba's latest flagship AI model for free before API access becomes the main way to use it

K2S

38,787 просмотров • 3 месяцев назад

Qwen3.8 Max UD-Q1_0 4× DGX Spark + llama.cpp/MTP3 53,448 tokens · 1h54m44s · 7.76 tok/s NVIDIA AI Qwen Qwen Developers

Qwen3.8 Max UD-Q1_0 4× DGX Spark + llama.cpp/MTP3 53,448 tokens · 1h54m44s · 7.76 tok/s NVIDIA AI Qwen Qwen Developers

Wësche

15,149 просмотров • 2 месяцев назад

Open-weight Qwen 3.8 2.4T built a Call of Duty clone in one prompt 🪖 Qwen released the 2.4T Max weights, so we rented a B200 cluster and asked the model to make a Call of Duty clone Output: ~1.1M tokens · 5 hours · one prompt Almost no one can run 2.4T at home, luckily there is a 27B version that runs on a 16GB MacBook Air with our Atomic Dynamic quants

Open-weight Qwen 3.8 2.4T built a Call of Duty clone in one prompt 🪖 Qwen released the 2.4T Max weights, so we rented a B200 cluster and asked the model to make a Call of Duty clone Output: ~1.1M tokens · 5 hours · one prompt Almost no one can run 2.4T at home, luckily there is a 27B version that runs on a 16GB MacBook Air with our Atomic Dynamic quants

atomic.chat

46,359 просмотров • 2 месяцев назад

Qwen-3.8 max 实战测评来啦！体验下来：参数量 2.4T，同任务下，速度上要比kimi-k3快很多，甚至有十倍的差距，对需要频繁交互的探索性任务比较友好，性价比较高这次也支持了多模态，上下文长度支持1M 前端能力从测评任务可以看到整体上还是不错的，与其他SOTA模型不相上下目前，Qwen3.8-Max 预览版模型（Qwen3.8-Max-Preview）已首发上线阿里 Token Plan、Qoder 及 QoderWork。当前 3.8看是限时一折，夜间更是0.2折，感兴趣的朋友可以试试且 Qwen3.8 正式版很快将发布并开源，正在以“天”为单位持续进化，今天意外发现测评的效果比昨天发布的版本要好很多，值得长期关注 Qwen3.8-Max-Preview 相比上一代 Qwen3.7-Max 提升明显；面对 KIMI-K3、GLM-5.2 和 Opus-4.8-Max 时也有一定优势，尤其擅长 Cowork 类 Agent 任务，但对 Fable-5-Xhigh 仍处于下风，很期待后续优化。评论附办公类测评👇

Qwen-3.8 max 实战测评来啦！体验下来：参数量 2.4T，同任务下，速度上要比kimi-k3快很多，甚至有十倍的差距，对需要频繁交互的探索性任务比较友好，性价比较高这次也支持了多模态，上下文长度支持1M 前端能力从测评任务可以看到整体上还是不错的，与其他SOTA模型不相上下目前，Qwen3.8-Max 预览版模型（Qwen3.8-Max-Preview）已首发上线阿里 Token Plan、Qoder 及 QoderWork。当前 3.8看是限时一折，夜间更是0.2折，感兴趣的朋友可以试试且 Qwen3.8 正式版很快将发布并开源，正在以“天”为单位持续进化，今天意外发现测评的效果比昨天发布的版本要好很多，值得长期关注 Qwen3.8-Max-Preview 相比上一代 Qwen3.7-Max 提升明显；面对 KIMI-K3、GLM-5.2 和 Opus-4.8-Max 时也有一定优势，尤其擅长 Cowork 类 Agent 任务，但对 Fable-5-Xhigh 仍处于下风，很期待后续优化。评论附办公类测评👇

岚叔

105,072 просмотров • 3 месяцев назад

Qwen3.7 Max Preview subway FPS test result

Qwen3.7 Max Preview subway FPS test result

BijanBowen

49,786 просмотров • 5 месяцев назад

🔥 K3 杀疯了，阿里坐不住了！今天，阿里千问正式预告 Qwen3.8 —— 参数量直接拉到 2.4T，官方放话："可能是除了 Fable 5 外最强大的模型"。还没正式发布，但 Qwen3.8-Max-Preview 已经抢先上线 Qoder、QoderWork 和千问 PC 端，限时 1 折调用，夜间甚至 0.2 折。内部测评显示，在全栈开发、数据分析、Office 办公等复杂长程任务上，它已经匹敌世界前沿模型。更夸张的是，我用它 1:1 复刻了《坦克大战》—— 碰撞、草地、道具、特种敌人，一句提示词全搞定。之前用 Fable 5 做过同样的 demo，这次 Qwen3.8 已经能和它扳手腕了。正式版即将开源，以"天"为单位进化中。 #Qwen3.8 #AI大模型

🔥 K3 杀疯了，阿里坐不住了！今天，阿里千问正式预告 Qwen3.8 —— 参数量直接拉到 2.4T，官方放话："可能是除了 Fable 5 外最强大的模型"。还没正式发布，但 Qwen3.8-Max-Preview 已经抢先上线 Qoder、QoderWork 和千问 PC 端，限时 1 折调用，夜间甚至 0.2 折。内部测评显示，在全栈开发、数据分析、Office 办公等复杂长程任务上，它已经匹敌世界前沿模型。更夸张的是，我用它 1:1 复刻了《坦克大战》—— 碰撞、草地、道具、特种敌人，一句提示词全搞定。之前用 Fable 5 做过同样的 demo，这次 Qwen3.8 已经能和它扳手腕了。正式版即将开源，以"天"为单位进化中。 #Qwen3.8 #AI大模型

OneHopeA9

43,142 просмотров • 3 месяцев назад

Chinese labs are now competing with each other to build the strongest open-weight model. > DeepSeek started the domino effect > Qwen kept improving across coding, reasoning and long context > GLM, MiniMax and Moonshot joined the race > Kimi K3 pushed it to 2.8T parameters and a 1M context window > now Alibaba is answering with Qwen3.8 > 2.4T parameters > full open-weight release coming soon > Alibaba says it is second only to Fable 5 Qwen3.8 may be the model that brings open-weight AI closest to Claude Fable 5.

Chinese labs are now competing with each other to build the strongest open-weight model. > DeepSeek started the domino effect > Qwen kept improving across coding, reasoning and long context > GLM, MiniMax and Moonshot joined the race > Kimi K3 pushed it to 2.8T parameters and a 1M context window > now Alibaba is answering with Qwen3.8 > 2.4T parameters > full open-weight release coming soon > Alibaba says it is second only to Fable 5 Qwen3.8 may be the model that brings open-weight AI closest to Claude Fable 5.

Tornado guy

65,279 просмотров • 3 месяцев назад