Home
Tools
Models
App
Prompts
News
Topics
Submit for inclusion
English
English 日本語 한국어 Português español Deutsch Русский Français 繁體中文 简体中文
option
Home Close
Home
Tools
Category Ranking Generate Cover Generate Image Skill
Models
Large Language Model Multimodal Model
App
Prompts
Prompts Image Prompts
News
News Flash News
Topics
Submit for inclusion
Submit App Submit Tool
Create an account Sign In
Change Language English Change Language
Settings
English EN 日本語 JA 한국어 KO Português PT español ES Deutsch DE Русский RU Français FR 繁體中文 ZH-TW 简体中文 ZH-CN
Home
Flash News
Content
FrankRoberts
FrankRoberts
September 22, 2026

Alibaba announced at CloudCon 2026 that Qwen4 is in training with future models scaling to 500B-1T parameters. Qwen3.8-Max achieved top rankings and demonstrated Recursive Self-Improvement by independently iterating for a month. The company highlighted Qwen3.8-Flash reducing training costs by 90% and Wan3.0 video generation. Open-source Qwen models have surpassed 3 billion downloads, with major enterprises like Airbnb and Perplexity adopting the technology for agents and AI assistants.

0
0
Share Click to copy the link
Comments (0)
0/300
Submit
Author ScottJohnson ScottJohnson September 23, 2026
Qualcomm unveiled Snapdragon 8 Elite Gen6 and 8 Elite Extreme Gen6 processors at its annual summit, prioritizing on-device AI. The Extreme version supports 30-billion parameter MoE models with dynamic activation for efficiency, while both chips feature enhanced imaging, 8K video, and noise-isolating voice tech. Motorola plans to launch the Signature 27 flagship with the Extreme chip this year, reinforcing smartphones as key devices for AI adoption.
0
0
Author BruceMitchell BruceMitchell September 23, 2026
Cisco Talos warns of ClosedQuorum, a novel Windows 11 malware that uses AI models like Gemini and DeepSeek for autonomous decision-making. Unlike traditional malware requiring human commands, it independently selects attack vectors such as credential theft and crypto extraction via a voting system. This first public AI-delegated C2 implant operates without human intervention, enhancing speed and scalability. Even if command servers go offline, the malware continues attacks based on local reconnaissance. Talos urges users to update patches and enable multi-factor authentication to mitigate risks from this new class of AI-driven threats.
0
0
/live/7573
Author StevenAllen StevenAllen September 23, 2026
ByteDance updated DouBao Work with Goal and Plan modes. Goal mode verifies complex tasks against standards, auto-reworking until met. Plan mode generates editable Markdown plans for user confirmation before execution, enhancing control and reducing resource use. New features include a task queue for continuous instructions, Markdown editing, and auto-switching dark mode. These updates shift AI office assistants from one-click execution to controlled, verifiable processes, improving productization.
0
0
/live/7572
Author BrianThomas BrianThomas September 23, 2026
Shokz unveiled its first AI-enabled earphone OpenFit2AI at the 2026 Yunqi Conference in Hangzhou integrating Alibaba's Qwen large model. The device features long-duration recording AI meeting notes and real-time translation optimized for finance and legal sectors with 19 industry templates. A new AI Lab introduces four features: intelligent conversation schedule management voice notes and a message assistant that extracts key content from long audio. This launch highlights deep industrial application of large models in audio wearables enhancing productivity by reducing phone usage through voice commands for weather queries info retrieval and schedule creation.
0
0
/live/7569
Author JonathanDavis JonathanDavis September 23, 2026
Volc Engine launched Draft mode for Seedance 2.5 API, splitting video generation into a low-cost 480P preview and 1080P final output. This two-stage process saves up to 77% cost and doubles speed while maintaining structural consistency through secondary reasoning. Users refine prompts based on the sample before generating native 1080P video, which offers richer details and natural lighting compared to traditional upscaling. Early testers from Runway, Magnific, and Artlist praised the mode for its speed, stability, and superior quality over standard upscaling methods.
0
0
/live/7568
OR