New image model form Alibaba: Z-Image
* just 6B
* fully self-attentional ("single stream") MM-DiT
* significant data annotation, augmentation, curation
distilled few-step "turbo" model released
paper: github.com/Tongyi-MAI/Z...
gh: github.com/Tongyi-MAI/Z...
hf: huggingface.co/Comfy-Org/z_...
* just 6B
* fully self-attentional ("single stream") MM-DiT
* significant data annotation, augmentation, curation
distilled few-step "turbo" model released
paper: github.com/Tongyi-MAI/Z...
gh: github.com/Tongyi-MAI/Z...
hf: huggingface.co/Comfy-Org/z_...
November 26, 2025 at 10:26 PM
New image model form Alibaba: Z-Image
* just 6B
* fully self-attentional ("single stream") MM-DiT
* significant data annotation, augmentation, curation
distilled few-step "turbo" model released
paper: github.com/Tongyi-MAI/Z...
gh: github.com/Tongyi-MAI/Z...
hf: huggingface.co/Comfy-Org/z_...
* just 6B
* fully self-attentional ("single stream") MM-DiT
* significant data annotation, augmentation, curation
distilled few-step "turbo" model released
paper: github.com/Tongyi-MAI/Z...
gh: github.com/Tongyi-MAI/Z...
hf: huggingface.co/Comfy-Org/z_...
Me, sitting in the doctor’s office yet again, looking at a poster: Huh. I wonder how you pronounce “tóngyì” (同意) in Mandarin-
Me:
Me:
Me: you know what
Me:
Me:
Me: you know what
December 9, 2024 at 7:08 AM
Me, sitting in the doctor’s office yet again, looking at a poster: Huh. I wonder how you pronounce “tóngyì” (同意) in Mandarin-
Me:
Me:
Me: you know what
Me:
Me:
Me: you know what
Alibaba launched Tongyi DeepResearch, the open-source web agent to achieve similar performance as OpenAI's Deep Research with only 30B (Activated 3B) parameters!
September 16, 2025 at 11:31 PM
Alibaba launched Tongyi DeepResearch, the open-source web agent to achieve similar performance as OpenAI's Deep Research with only 30B (Activated 3B) parameters!
Alibaba releases a powerful image generator, Z-Image, with 6B parameters in three variants that shows highly competitive performance against other leading models, while achieving state-of-the-art results among open-source models.
github.com/Tongyi-MAI/Z...
github.com/Tongyi-MAI/Z...
GitHub - Tongyi-MAI/Z-Image
Contribute to Tongyi-MAI/Z-Image development by creating an account on GitHub.
github.com
November 28, 2025 at 4:13 PM
Alibaba releases a powerful image generator, Z-Image, with 6B parameters in three variants that shows highly competitive performance against other leading models, while achieving state-of-the-art results among open-source models.
github.com/Tongyi-MAI/Z...
github.com/Tongyi-MAI/Z...
Qwen Tongyi Deep Research
a 32B model that beats other SOTA deep research on many benchmarks
tongyi-agent.github.io/blog/introdu...
a 32B model that beats other SOTA deep research on many benchmarks
tongyi-agent.github.io/blog/introdu...
September 16, 2025 at 10:37 PM
Qwen Tongyi Deep Research
a 32B model that beats other SOTA deep research on many benchmarks
tongyi-agent.github.io/blog/introdu...
a 32B model that beats other SOTA deep research on many benchmarks
tongyi-agent.github.io/blog/introdu...
Alibaba's Tongyi DeepResearch Technical Report
Dive deep into the technology and insights behind our 30B (A3B) open-source web agent that achieves SOTA performance: 32.9 on Humanity's Last Exam, 43.4 on BrowseComp, and 46.7 on BrowseComp-ZH.
Dive deep into the technology and insights behind our 30B (A3B) open-source web agent that achieves SOTA performance: 32.9 on Humanity's Last Exam, 43.4 on BrowseComp, and 46.7 on BrowseComp-ZH.
October 29, 2025 at 2:38 PM
Alibaba's Tongyi DeepResearch Technical Report
Dive deep into the technology and insights behind our 30B (A3B) open-source web agent that achieves SOTA performance: 32.9 on Humanity's Last Exam, 43.4 on BrowseComp, and 46.7 on BrowseComp-ZH.
Dive deep into the technology and insights behind our 30B (A3B) open-source web agent that achieves SOTA performance: 32.9 on Humanity's Last Exam, 43.4 on BrowseComp, and 46.7 on BrowseComp-ZH.
Alibaba released a paper on their Tongyi DeepResearch. They incorporated Agentic Continual Pre-training (Agentic CPT) into the deep research agents training pipeline to build powerful agentic foundational models.
Paper: Scaling Agents via Continual Pre-training
arxiv.org/abs/2509.13310
Paper: Scaling Agents via Continual Pre-training
arxiv.org/abs/2509.13310
September 18, 2025 at 12:49 AM
Alibaba released a paper on their Tongyi DeepResearch. They incorporated Agentic Continual Pre-training (Agentic CPT) into the deep research agents training pipeline to build powerful agentic foundational models.
Paper: Scaling Agents via Continual Pre-training
arxiv.org/abs/2509.13310
Paper: Scaling Agents via Continual Pre-training
arxiv.org/abs/2509.13310
⚠️ Update on Deep Research Tulu (DR Tulu), our post-training recipe for deep research agents: we’re releasing an upgraded version of our example agent, DR Tulu-8B (RL), that matches or beats systems like Gemini 3 Pro & Tongyi DeepResearch-30B-A3B on core benchmarks. 🧵
November 25, 2025 at 7:37 PM
⚠️ Update on Deep Research Tulu (DR Tulu), our post-training recipe for deep research agents: we’re releasing an upgraded version of our example agent, DR Tulu-8B (RL), that matches or beats systems like Gemini 3 Pro & Tongyi DeepResearch-30B-A3B on core benchmarks. 🧵
wo tongyi
January 14, 2025 at 11:47 AM
wo tongyi
Loaded up good ole Tōngyì Qiānwèn 3.8 27B on my GTX4090 with this and it immediately sounds like Astra and is prioritizing way better on its own. wtflmao
I asked Astra to put this work somewhere other people could use it, so now it's a repo. mostly LLM-written and not well human-curated; I endorse it, but the null results are in there too.
github.com/rmc3/recipro...
github.com/rmc3/recipro...
October 3, 2026 at 2:08 PM
Loaded up good ole Tōngyì Qiānwèn 3.8 27B on my GTX4090 with this and it immediately sounds like Astra and is prioritizing way better on its own. wtflmao
Alibaba's InspireMusic – an open-source music generation toolkit from Tongyi Lab, designed as an all-in-one AIGC toolkit for music, song, and audio creation.
🎵 GitHub Repository: github.com/FunAudioLLM/...
🎶 Online Experience: iris2c.github.io/InspireMusic/
🎵 GitHub Repository: github.com/FunAudioLLM/...
🎶 Online Experience: iris2c.github.io/InspireMusic/
February 11, 2025 at 6:47 AM
Alibaba's InspireMusic – an open-source music generation toolkit from Tongyi Lab, designed as an all-in-one AIGC toolkit for music, song, and audio creation.
🎵 GitHub Repository: github.com/FunAudioLLM/...
🎶 Online Experience: iris2c.github.io/InspireMusic/
🎵 GitHub Repository: github.com/FunAudioLLM/...
🎶 Online Experience: iris2c.github.io/InspireMusic/
Le renseignement taïwanais (NSB) exhorte à faire preuve de vigilance à l'égard des Chatbots IA chinois, en raison des failles de sécurité potentielles et de la propagation de « désinformation » mises en évidence lors d'inspections de Deepseek, Doubao, Yiyan, Tongyi et Yuanbao
November 18, 2025 at 6:40 AM
Le renseignement taïwanais (NSB) exhorte à faire preuve de vigilance à l'égard des Chatbots IA chinois, en raison des failles de sécurité potentielles et de la propagation de « désinformation » mises en évidence lors d'inspections de Deepseek, Doubao, Yiyan, Tongyi et Yuanbao
github.com/Tongyi-MAI/Z...
Z-Image is aa image generation model with 6B parameters 😮
Z-Image is aa image generation model with 6B parameters 😮
GitHub - Tongyi-MAI/Z-Image
Contribute to Tongyi-MAI/Z-Image development by creating an account on GitHub.
github.com
December 7, 2025 at 9:54 AM
github.com/Tongyi-MAI/Z...
Z-Image is aa image generation model with 6B parameters 😮
Z-Image is aa image generation model with 6B parameters 😮
Tongyi-MAI/Z-Image: a powerful and highly efficient image generation model with 6B parameters github.com/Tongyi-MAI/Z...
GitHub - Tongyi-MAI/Z-Image
Contribute to Tongyi-MAI/Z-Image development by creating an account on GitHub.
github.com
November 29, 2025 at 9:35 PM
Tongyi-MAI/Z-Image: a powerful and highly efficient image generation model with 6B parameters github.com/Tongyi-MAI/Z...
cuando leí The Imperial Uncle siempre decían que Liu Tongyi se tenia que cambiar de tus túnicas de la corte y pues no sabia muuuuuy bien qué imaginarme pero Follow Your Heart me enseñó que tenían uniforme para la corte, jajaja y ahora siempre lo imagino con su gorrito chistoso:
November 15, 2024 at 11:20 PM
cuando leí The Imperial Uncle siempre decían que Liu Tongyi se tenia que cambiar de tus túnicas de la corte y pues no sabia muuuuuy bien qué imaginarme pero Follow Your Heart me enseñó que tenían uniforme para la corte, jajaja y ahora siempre lo imagino con su gorrito chistoso:
i can be your cartesian demon bb
July 15, 2026 at 7:44 PM
i can be your cartesian demon bb
Tongyi DeepResearch is an open-source agent that performs multi-step reasoning with minimal human input. With an innovative framework and just 3.3 billion parameters activated per token, it sets a new AI standard and boosts productivity in academia and beyond. https://arxiv.org/abs/2510.24701
Tongyi DeepResearch Technical Report
ArXiv link for Tongyi DeepResearch Technical Report
arxiv.org
May 19, 2026 at 8:50 PM
Tongyi DeepResearch is an open-source agent that performs multi-step reasoning with minimal human input. With an innovative framework and just 3.3 billion parameters activated per token, it sets a new AI standard and boosts productivity in academia and beyond. https://arxiv.org/abs/2510.24701
An...interesting translation of Chunqiu Gongyang jingzhuan tongyi 春秋公羊經傳通義 (lit. Comprehensive Meanings of the Classic and Tradition of the Gongyang Commentary to the Annals) by Kong Guangsen 孔廣森 (1753-1786). #chineseclassics #confucianism #china #chinese #philosophy
October 17, 2025 at 11:29 AM
An...interesting translation of Chunqiu Gongyang jingzhuan tongyi 春秋公羊經傳通義 (lit. Comprehensive Meanings of the Classic and Tradition of the Gongyang Commentary to the Annals) by Kong Guangsen 孔廣森 (1753-1786). #chineseclassics #confucianism #china #chinese #philosophy
Alibaba Qwenチームの主任研究者ら主要メンバーが突如退職
https://gigazine.net/news/20260305-alibaba-step-down/
https://gigazine.net/news/20260305-alibaba-step-down/
Alibaba Qwenチームの主任研究者ら主要メンバーが突如退職
AlibabaのAI研究チームであるQwen(Tongyi Lab)から、中心的な技術的リーダーを含む複数の主要メンバーが突然離脱したことが報じられています。
gigazine.net
March 5, 2026 at 4:41 AM
Alibaba Qwenチームの主任研究者ら主要メンバーが突如退職
https://gigazine.net/news/20260305-alibaba-step-down/
https://gigazine.net/news/20260305-alibaba-step-down/
The image(s) was made with Grok and turned into a video using Wan 2.1, an AI-powered image-to-video piece of software developed by Alibaba-owned Tongyi Lab. www.wan-ai.org
April 23, 2025 at 1:38 AM
The image(s) was made with Grok and turned into a video using Wan 2.1, an AI-powered image-to-video piece of software developed by Alibaba-owned Tongyi Lab. www.wan-ai.org
Alibaba demos R1-Omni, an open-source model that its Tongyi Lab researchers say can read emotions, building on HumanOmni, in a bid to compete with OpenAI (Luz Ding/Bloomberg)
Main Link | Techmeme Permalink
Main Link | Techmeme Permalink
March 12, 2025 at 11:46 AM
Alibaba demos R1-Omni, an open-source model that its Tongyi Lab researchers say can read emotions, building on HumanOmni, in a bid to compete with OpenAI (Luz Ding/Bloomberg)
Main Link | Techmeme Permalink
Main Link | Techmeme Permalink
Sources: Alibaba is set to lead a $300M round in AI model testing startup UniPat AI at a $2.5B valuation; UniPat founder Li Kuan worked at Alibaba's Tongyi Lab (Bloomberg)
Main Link | Techmeme Permalink
Main Link | Techmeme Permalink
September 10, 2026 at 4:06 AM
Sources: Alibaba is set to lead a $300M round in AI model testing startup UniPat AI at a $2.5B valuation; UniPat founder Li Kuan worked at Alibaba's Tongyi Lab (Bloomberg)
Main Link | Techmeme Permalink
Main Link | Techmeme Permalink
the-deepseek-moment-for-ai-agents-is-here-meet-alibabas-open-source-tongyi https://venturebeat.com/ai/the-deepseek-moment-for-ai-agents-is-here-meet-alibabas-open-source-tongyi (another Deep Research) #AI #OpenSource
September 19, 2025 at 4:06 PM
the-deepseek-moment-for-ai-agents-is-here-meet-alibabas-open-source-tongyi https://venturebeat.com/ai/the-deepseek-moment-for-ai-agents-is-here-meet-alibabas-open-source-tongyi (another Deep Research) #AI #OpenSource
Alibaba's Tongyi Qwen Team Releases QwQ-32B Model Alibaba's Tongyi Qwen team releases open-source QwQ-32B model with 32B parameters, matching DeepSeek-R1 model's 671B parameters in performance. The team highlights reinforcement learning (RL) as a potential method to enhance model…
March 6, 2025 at 5:28 AM
Alibaba's Tongyi Qwen Team Releases QwQ-32B Model Alibaba's Tongyi Qwen team releases open-source QwQ-32B model with 32B parameters, matching DeepSeek-R1 model's 671B parameters in performance. The team highlights reinforcement learning (RL) as a potential method to enhance model…