正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Qwen3-Next (thinking & non-thinking) are now live in BF16 at Hyperbolic! Qwen3-Next is a huge efficiency leap: - 80B MoE with just 3B active params - 10x cheaper to train vs Qwen3-32B - 10x inference throughput for >32K tokens Proud to be a launch partner with Qwen - kudos... show more

Yuchen Jin

61,835 subscribers

84,342 次观看 • 10 个月前 •via X (Twitter)

科学技术新闻政治教育

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

🤯 Wow! In one prompt Qwen3-Coder-Next generated a fully working flappy birds game in HTML. (0:05) Claude Code with Qwen3-Coder-Next (0:26) Shows the game running Run it fully locally: ollama pull qwen3-coder-next Ollama's cloud if you can't run it locally: ollama pull qwen3-coder-next:cloud Try launching it with Claude Code using ollama launch (link to play 🧵) So cool! Qwen Tongyi Lab Junyang Lin

🤯 Wow! In one prompt Qwen3-Coder-Next generated a fully working flappy birds game in HTML. (0:05) Claude Code with Qwen3-Coder-Next (0:26) Shows the game running Run it fully locally: ollama pull qwen3-coder-next Ollama's cloud if you can't run it locally: ollama pull qwen3-coder-next:cloud Try launching it with Claude Code using ollama launch (link to play 🧵) So cool! Qwen Tongyi Lab Junyang Lin

ollama

117,339 次观看 • 5 个月前

1/4Introducing Qwen3-Coder-Next: Our latest open-weights model designed specifically to power the next generation of autonomous Coding Agents. Built on Qwen3-Next, this model is engineered to handle complex, long-horizon programming tasks with unprecedented efficiency. High-performance agentic intelligence is now in your hands.

1/4Introducing Qwen3-Coder-Next: Our latest open-weights model designed specifically to power the next generation of autonomous Coding Agents. Built on Qwen3-Next, this model is engineered to handle complex, long-horizon programming tasks with unprecedented efficiency. High-performance agentic intelligence is now in your hands.

Tongyi Lab

212,645 次观看 • 5 个月前

Launched by popular demand. Qwen3-32B is now live on GroqCloud. Supports over 100 languages and dialects with a 131k context window. It is not incremental. It is a leap. Only Groq delivers this level of context with real-time speed. Demo video below to spark your next build. GitHub link in the comments. ⬇️

Launched by popular demand. Qwen3-32B is now live on GroqCloud. Supports over 100 languages and dialects with a 131k context window. It is not incremental. It is a leap. Only Groq delivers this level of context with real-time speed. Demo video below to spark your next build. GitHub link in the comments. ⬇️

Groq Inc

26,498 次观看 • 1 年前

icymi Hugging Face dropped a computer use agent last week 🔥 built on various Qwen3-VL models & E2B sandbox, ask the app to do anything 🙌🏻 it exposes each thinking step to you, try different models with a neat UI🤩

icymi Hugging Face dropped a computer use agent last week 🔥 built on various Qwen3-VL models & E2B sandbox, ask the app to do anything 🙌🏻 it exposes each thinking step to you, try different models with a neat UI🤩

merve

16,066 次观看 • 7 个月前

Over the past few days, I’ve been actively testing Qwen3-Coder-Next, and I’ve built quite a few fun and creative mini-games with it. Its frontend capabilities are genuinely impressive. Because it’s based on the Qwen3-Next model, the speed is outstanding. In fact, we managed to build an Angry Birds–style game in just 45 seconds, which really pushed the upper bound of what I thought was possible in terms of efficiency. It’s also been extremely useful in my day-to-day work. I used it to read through the Qwen3-TTS GitHub repository, quickly write a Gradio app, and deploy it on ModelScope. The overall performance and accuracy were excellent—especially considering that Qwen3-TTS is not a trivial application at all. Overall, this is a model I genuinely enjoy using. I warmly welcome any feedback, and bad cases are especially appreciated—they’re incredibly helpful for us as we continue to optimize and improve the model.

Over the past few days, I’ve been actively testing Qwen3-Coder-Next, and I’ve built quite a few fun and creative mini-games with it. Its frontend capabilities are genuinely impressive. Because it’s based on the Qwen3-Next model, the speed is outstanding. In fact, we managed to build an Angry Birds–style game in just 45 seconds, which really pushed the upper bound of what I thought was possible in terms of efficiency. It’s also been extremely useful in my day-to-day work. I used it to read through the Qwen3-TTS GitHub repository, quickly write a Gradio app, and deploy it on ModelScope. The overall performance and accuracy were excellent—especially considering that Qwen3-TTS is not a trivial application at all. Overall, this is a model I genuinely enjoy using. I warmly welcome any feedback, and bad cases are especially appreciated—they’re incredibly helpful for us as we continue to optimize and improve the model.

Chen Cheng

18,396 次观看 • 5 个月前

I seriously cannot believe this is a 0.6B LLM! 🤯 Qwen just released Qwen3, a series of hybrid reasoning models that allow you to control how much "thinking" the model does for a given task. They can even run locally in your browser on WebGPU with 🤗 Transformers.js!

I seriously cannot believe this is a 0.6B LLM! 🤯 Qwen just released Qwen3, a series of hybrid reasoning models that allow you to control how much "thinking" the model does for a given task. They can even run locally in your browser on WebGPU with 🤗 Transformers.js!

Xenova

74,954 次观看 • 1 年前

Jan v1 is now more reliable when handling searches. It's an open-source alternative to Perplexity Pro, and it now avoids infinite loops with Jan v1 2509. Performance updates: - Gains on reasoning & creativity benchmarks - Small drop on SimpleQA (91.1 -> 90.7), still competitive with Perplexity Pro The updated version is live on Jan Hub. You can run the model in Jan, llama.cpp, or vLLM. Use the model: - Model: - GGUF: Shoutout to the Qwen team for Qwen3-4B Thinking and Georgi Gerganov for llama.cpp.

Jan v1 is now more reliable when handling searches. It's an open-source alternative to Perplexity Pro, and it now avoids infinite loops with Jan v1 2509. Performance updates: - Gains on reasoning & creativity benchmarks - Small drop on SimpleQA (91.1 -> 90.7), still competitive with Perplexity Pro The updated version is live on Jan Hub. You can run the model in Jan, llama.cpp, or vLLM. Use the model: - Model: - GGUF: Shoutout to the Qwen team for Qwen3-4B Thinking and Georgi Gerganov for llama.cpp.

👋 Jan

17,759 次观看 • 10 个月前

Introducing Jan-v1: 4B model for web search, an open-source alternative to Perplexity Pro. In our evals, Jan v1 delivers 91% SimpleQA accuracy, slightly outperforming Perplexity Pro while running fully locally. Use cases: - Web search - Deep Research Built on the new version of Qwen's Qwen3-4B-Thinking (up to 256k context length), fine-tuned for reasoning and tool use in Jan. You can run the model in Jan, llama.cpp, or vLLM. To enable search in Jan, go to Settings → Experimental Features → On, then Settings → MCP Servers → enable a search-related MCP such as Serper. Use the model: - Jan-v1-4B: - Jan-v1-4B-GGUF: Credit to the Qwen team for Qwen3 4B Thinking & Georgi Gerganov for llama.cpp.

Introducing Jan-v1: 4B model for web search, an open-source alternative to Perplexity Pro. In our evals, Jan v1 delivers 91% SimpleQA accuracy, slightly outperforming Perplexity Pro while running fully locally. Use cases: - Web search - Deep Research Built on the new version of Qwen's Qwen3-4B-Thinking (up to 256k context length), fine-tuned for reasoning and tool use in Jan. You can run the model in Jan, llama.cpp, or vLLM. To enable search in Jan, go to Settings → Experimental Features → On, then Settings → MCP Servers → enable a search-related MCP such as Serper. Use the model: - Jan-v1-4B: - Jan-v1-4B-GGUF: Credit to the Qwen team for Qwen3 4B Thinking & Georgi Gerganov for llama.cpp.

👋 Jan

690,311 次观看 • 11 个月前

timelapse #85 (27.5 hrs): - currently cant rely on any other coding models except grok code fast 1 + grok 4 fast (for complex reasoning grok 4 fast is 20 cents for 1M tokens) - wrote qwen3-next trainer entirely from scratch to make it more managable - each piece completely done by grok-code-fast-1 in cursor as it seems to handle this task pretty well without the grok 4 fast reasoning - take on smaller problems and complete them quickly (makes it easier with 400 toks/sec over the api) - got distributed fp8 qwen3-next trainer running at 0.8 seconds per step on 8xH100s (still need to finish checkpoint loading logic) - perfect timing as the fp8 version of qwen3-next drops as im writing this - ill be in LA in 2 days (will visit SF mid way through as well) - 12.5% margarita - steak dinner with family - gained intuition on FlashAttention in very long context settings - caught up w/ Kearm h/eng and Arnie Ramesh

timelapse #85 (27.5 hrs): - currently cant rely on any other coding models except grok code fast 1 + grok 4 fast (for complex reasoning grok 4 fast is 20 cents for 1M tokens) - wrote qwen3-next trainer entirely from scratch to make it more managable - each piece completely done by grok-code-fast-1 in cursor as it seems to handle this task pretty well without the grok 4 fast reasoning - take on smaller problems and complete them quickly (makes it easier with 400 toks/sec over the api) - got distributed fp8 qwen3-next trainer running at 0.8 seconds per step on 8xH100s (still need to finish checkpoint loading logic) - perfect timing as the fp8 version of qwen3-next drops as im writing this - ill be in LA in 2 days (will visit SF mid way through as well) - 12.5% margarita - steak dinner with family - gained intuition on FlashAttention in very long context settings - caught up w/ Kearm h/eng and Arnie Ramesh

Elliot Arledge

283,820 次观看 • 9 个月前

They are not there, but if you have a beefy mac, Qwen3-Coder-Next has some applications due to its speed and decent tool calling capabilities. This video is in real time. (Also plenty of local models for STT etc.). They are not to be completely dismissed!

They are not there, but if you have a beefy mac, Qwen3-Coder-Next has some applications due to its speed and decent tool calling capabilities. This video is in real time. (Also plenty of local models for STT etc.). They are not to be completely dismissed!

Armin Ronacher ⇌

36,110 次观看 • 3 个月前

Recipe to post-train Qwen3 1.7B into a DeepResearch model What does it mean for something small to think deeply? Meet Lucy, a post‑trained Qwen3‑1.7B as a DeepResearch model based on will brown's verifiers. Primary Rule-based Rewards: - Answer correctness We check whether the final response literally contains the ground-truth answer. This substring match is cheap and avoids calling a larger LLM judge. - Visit/search ratio If the agent visits at least as many pages as it issues search queries, it receives ((visit_search_ratio - 1) / 4) ** 0.25. If it searches more than it visits, the score is -0.5. Format / Anti Reward-Hacking Rewards: - Tool execution success Each API call that returns without an error counts. The reward is (successful_calls * unique_tools_used) / total_call_attempts. - Thinking efficiency A skew-normal penalty centered at 70 tokens discourages endless chain-of-thought between tool calling while still allowing enough tokens for planning. This is how Qwen3 1.7B learned to search, visit, and synthesize information. Small models can do deep research too!

Recipe to post-train Qwen3 1.7B into a DeepResearch model What does it mean for something small to think deeply? Meet Lucy, a post‑trained Qwen3‑1.7B as a DeepResearch model based on will brown's verifiers. Primary Rule-based Rewards: - Answer correctness We check whether the final response literally contains the ground-truth answer. This substring match is cheap and avoids calling a larger LLM judge. - Visit/search ratio If the agent visits at least as many pages as it issues search queries, it receives ((visit_search_ratio - 1) / 4) ** 0.25. If it searches more than it visits, the score is -0.5. Format / Anti Reward-Hacking Rewards: - Tool execution success Each API call that returns without an error counts. The reward is (successful_calls * unique_tools_used) / total_call_attempts. - Thinking efficiency A skew-normal penalty centered at 70 tokens discourages endless chain-of-thought between tool calling while still allowing enough tokens for planning. This is how Qwen3 1.7B learned to search, visit, and synthesize information. Small models can do deep research too!

Casper Hansen

39,684 次观看 • 11 个月前

🚀 VoxCPM 2 is live! 🎉 Another open-source AI #TTS model from China — and one that stands shoulder to shoulder with Qwen3-TTS, while bringing everything into a single unified model. After rapid iterations from V1 (zero-shot cloning) to V1.5 (long-form + fine-tuning), #VoxCPM has consistently pushed quality and usability forward. Now, VoxCPM 2 takes it further: 🔹30+ languages — truly global, truly local. 🔹Infinite voice design — type it, hear it, control it. From a whisper to a booming cinematic voice. 🔹Studio-grade audio — 48kHz ultra-high fidelity with emotional depth 🔹Diffusion-Autoregressive cloning — preserves more acoustic and emotional detail than token-based models like Qwen3-TTS 💡 Big shoutout to Grok — used your multi-image video magic for our launch demo. It’s scarily good at keeping visuals consistent across shots. Elon Elon Musk, this one’s for you. 😉 Check the demo & start cloning your dream voice: 🌐 Hugging Face Space: 🤗 Hugging Face Model: 🤖 ModelScope Model: 💻 GitHub： #TTS #AI #VoiceCloning #GrokImagine #ElonMusk #OpenBMB #VoxCPM

🚀 VoxCPM 2 is live! 🎉 Another open-source AI #TTS model from China — and one that stands shoulder to shoulder with Qwen3-TTS, while bringing everything into a single unified model. After rapid iterations from V1 (zero-shot cloning) to V1.5 (long-form + fine-tuning), #VoxCPM has consistently pushed quality and usability forward. Now, VoxCPM 2 takes it further: 🔹30+ languages — truly global, truly local. 🔹Infinite voice design — type it, hear it, control it. From a whisper to a booming cinematic voice. 🔹Studio-grade audio — 48kHz ultra-high fidelity with emotional depth 🔹Diffusion-Autoregressive cloning — preserves more acoustic and emotional detail than token-based models like Qwen3-TTS 💡 Big shoutout to Grok — used your multi-image video magic for our launch demo. It’s scarily good at keeping visuals consistent across shots. Elon Elon Musk, this one’s for you. 😉 Check the demo & start cloning your dream voice: 🌐 Hugging Face Space: 🤗 Hugging Face Model: 🤖 ModelScope Model: 💻 GitHub： #TTS #AI #VoiceCloning #GrokImagine #ElonMusk #OpenBMB #VoxCPM

OpenBMB

557,835 次观看 • 3 个月前

✨ Personal AI is the next computing platform. AI is shifting from something you access to something you build with, locally, at the edge, and across systems. We’re unlocking new possibilities for developers: • AMD Ryzen AI Halo, a local-first developer system, preorder starting in June, develop AI without limits on your desk • Gorgon Halo with up to 192GB unified memory, supporting 300B+ parameter models locally, run massive models locally We’re excited to partner with clem 🤗🤗, Co-founder and CEO of Hugging Face, to advance open-source AI for Ryzen AI. Our focus is seamless AI, from model to deployment. Cloud, edge, device. One continuum. Multi-agent systems, local inference at scale, open models as infrastructure. This is the next computing era 🚀

✨ Personal AI is the next computing platform. AI is shifting from something you access to something you build with, locally, at the edge, and across systems. We’re unlocking new possibilities for developers: • AMD Ryzen AI Halo, a local-first developer system, preorder starting in June, develop AI without limits on your desk • Gorgon Halo with up to 192GB unified memory, supporting 300B+ parameter models locally, run massive models locally We’re excited to partner with clem 🤗🤗, Co-founder and CEO of Hugging Face, to advance open-source AI for Ryzen AI. Our focus is seamless AI, from model to deployment. Cloud, edge, device. One continuum. Multi-agent systems, local inference at scale, open models as infrastructure. This is the next computing era 🚀

Jack Huynh

47,822 次观看 • 1 个月前

As AI labs race to train and deploy new frontier models, existing models become more affordable with better tokenomics. ✨ "Everybody's trying to get to the next frontier. And every time they get to the next frontier, the last generation AI tokens, the cost starts to decline about a factor of 10x every year," said NVIDIA CEO Jensen Huang in a recent keynote. Model optimization techniques such as speculative decoding and multi-token prediction, combined with inference serving platforms like NVIDIA Dynamo on NVIDIA Blackwell NVL72 systems, enable AI factories to boost throughput by 10x with one-tenth of the cost per token. Learn more about AI factory tokenomics ➡️

As AI labs race to train and deploy new frontier models, existing models become more affordable with better tokenomics. ✨ "Everybody's trying to get to the next frontier. And every time they get to the next frontier, the last generation AI tokens, the cost starts to decline about a factor of 10x every year," said NVIDIA CEO Jensen Huang in a recent keynote. Model optimization techniques such as speculative decoding and multi-token prediction, combined with inference serving platforms like NVIDIA Dynamo on NVIDIA Blackwell NVL72 systems, enable AI factories to boost throughput by 10x with one-tenth of the cost per token. Learn more about AI factory tokenomics ➡️

NVIDIA AI

16,053 次观看 • 5 个月前

Hugging Face CEO Clem Delangue predicts that a growing share of AI inference will run locally, on-device. It's faster. Cheaper. Private by default. And as open source AI moves from apps to machines, it could trigger “some sort of ChatGPT moment for robotics.” The next AI breakthrough will happen where intelligence becomes embodied, cheap, private, and open.

Hugging Face CEO Clem Delangue predicts that a growing share of AI inference will run locally, on-device. It's faster. Cheaper. Private by default. And as open source AI moves from apps to machines, it could trigger “some sort of ChatGPT moment for robotics.” The next AI breakthrough will happen where intelligence becomes embodied, cheap, private, and open.

vitrupo

37,764 次观看 • 1 年前

opencode making a pong game in vite+react using (4bit) qwen/qwen3-235b-a22b-2507 locally, served by lmstudio. It used like 130GB of RAM, 0 issues with tool calls. This is completely usable locally now. Whether it's at claude level or not, idk yet, but I've no doubt we'll be there very soon if not. Anthropic better run fast to the Saudis for help.

opencode making a pong game in vite+react using (4bit) qwen/qwen3-235b-a22b-2507 locally, served by lmstudio. It used like 130GB of RAM, 0 issues with tool calls. This is completely usable locally now. Whether it's at claude level or not, idk yet, but I've no doubt we'll be there very soon if not. Anthropic better run fast to the Saudis for help.

SIGKITTEN

30,470 次观看 • 11 个月前

Opening orders for Reachy Mini today, our open-source desktop robot for AI builders, starting at $299! Fully integrated with LeRobot & Hugging Face for the whole community to build AI apps for it (like this dancing one). We'll probably ship a first batch of a hundred this summer, first come first serve!

Opening orders for Reachy Mini today, our open-source desktop robot for AI builders, starting at $299! Fully integrated with LeRobot & Hugging Face for the whole community to build AI apps for it (like this dancing one). We'll probably ship a first batch of a hundred this summer, first come first serve!

clem 🤗

108,193 次观看 • 1 年前

𝗢𝗹𝗹𝗮𝗺𝗮 + 𝗖𝗹𝗮𝘂𝗱𝗲 𝗖𝗼𝗱𝗲 𝗷𝘂𝘀𝘁 𝗺𝗮𝗱𝗲 𝗔𝗜 𝗰𝗼𝗱𝗶𝗻𝗴 𝗳𝗿𝗲𝗲. 𝗔𝗻𝗱 𝗶𝘁 𝗿𝘂𝗻𝘀 𝗰𝗼𝗺𝗽𝗹𝗲𝘁𝗲𝗹𝘆 𝗹𝗼𝗰𝗮𝗹. No subscriptions. No limits. No data leaving your machine. Claude Code used to cost $3-$15 per million tokens. Now you can run it with free open-source models on your computer. January 16th, 2026: Ollama version 0.14.0 became compatible with Anthropic's messages API. That means Claude Code now works with any Ollama model. Zero ongoing costs. Your code never leaves your computer. Here's the setup: Download Ollama. Run `ollama pull qwen3-coder`. Install Claude Code. Set two environment variables to connect them. Then run: `claude --model qwen3-coder` Give it tasks in plain English: "Write a function to process images." "Debug why my API calls are failing." It reads your files. Makes changes. Tests the code. Shows you what it did. Top models: Qwen 3-Coder for everyday tasks. GPT-OSS 20B for complex projects. This is completely legal. Explicitly supported by both tools.

𝗢𝗹𝗹𝗮𝗺𝗮 + 𝗖𝗹𝗮𝘂𝗱𝗲 𝗖𝗼𝗱𝗲 𝗷𝘂𝘀𝘁 𝗺𝗮𝗱𝗲 𝗔𝗜 𝗰𝗼𝗱𝗶𝗻𝗴 𝗳𝗿𝗲𝗲. 𝗔𝗻𝗱 𝗶𝘁 𝗿𝘂𝗻𝘀 𝗰𝗼𝗺𝗽𝗹𝗲𝘁𝗲𝗹𝘆 𝗹𝗼𝗰𝗮𝗹. No subscriptions. No limits. No data leaving your machine. Claude Code used to cost $3-$15 per million tokens. Now you can run it with free open-source models on your computer. January 16th, 2026: Ollama version 0.14.0 became compatible with Anthropic's messages API. That means Claude Code now works with any Ollama model. Zero ongoing costs. Your code never leaves your computer. Here's the setup: Download Ollama. Run `ollama pull qwen3-coder`. Install Claude Code. Set two environment variables to connect them. Then run: `claude --model qwen3-coder` Give it tasks in plain English: "Write a function to process images." "Debug why my API calls are failing." It reads your files. Makes changes. Tests the code. Shows you what it did. Top models: Qwen 3-Coder for everyday tasks. GPT-OSS 20B for complex projects. This is completely legal. Explicitly supported by both tools.

Julian Goldie SEO

168,502 次观看 • 5 个月前