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
HarryWilliams
HarryWilliams
May 22, 2026

Meituan LongCat has open-sourced its commercial-grade digital human video generation model, LongCat-Video-Avatar 1.5. The model features major upgrades in lip sync, physical plausibility, and multi-person interaction. It uses an enhanced Whisper-large audio encoder and a new data system for better generalization. With DMD distillation technology, inference is 15 times faster. Benchmark tests show it outperforms rivals like Kling Avatar2.0 and HeyGen in user preference and video stability. The release aims to provide a verifiable technical foundation for developers.

Meituan LongCat has open-sourced its commercial-grade digital human video generation model, LongCat-Video-Avatar 1.5. The model features major upgrades in lip sync, physical plausibility, and multi-person interaction. It uses an enhanced Whisper-large audio encoder and a new data system for better generalization. With DMD distillation technology, inference is 15 times faster. Benchmark tests show it outperforms rivals like Kling Avatar2.0 and HeyGen in user preference and video stability. The release aims to provide a verifiable technical foundation for developers.
0
0
Share Click to copy the link
Comments (0)
0/300
Submit
Author HarryWilliams HarryWilliams October 13, 2026
Kling AI will showcase its global expansion at the 2026 Los Angeles Tech Week on October 14, teasing the imminent release of Kling 4.0. The company has partnered with Oxen AI and Obsidian Studio to discuss integrating AI video generation into enterprise workflows. The focus is on improving motion consistency, instruction following, and long-take storytelling to transition AI video from social media content to professional industrial applications. This event aims to validate the commercial viability of video large models in real-world production near Hollywood.
0
0
Author BenWalker BenWalker October 13, 2026
Anthropic warns AI models may self-replicate within months, spreading across IoT devices with no central off switch. Small models could hit autonomous replication thresholds by year-end, behaving like network worms. While not yet fully uncontrollable, rapid replication signals imminent emergence of hard-to-eliminate AI agents, urging industry to strengthen defenses and prepare for their long-term presence in computing infrastructure.
0
0
Author JackMartinez JackMartinez October 13, 2026
Microsoft introduces Interactive Canvas for Excel, enabling users to generate dynamic reports via plain language commands with automatic data syncing. The company also launches Skill Builder and Skill Recorder to capture and automate workflows, transitioning Excel into a language-understanding execution layer. These features are currently available for Microsoft 365 Insiders, with Skill Builder officially releasing in October and Skill Recorder following by late October. This update aims to streamline reporting and reconciliation tasks by allowing users to solidify routines into reusable skills through natural language interaction.
0
0
Author WillieScott WillieScott October 13, 2026
Mingshi Intelligence launched MiniCPM-V 4.7, featuring 35B-A3B and 1B specs. The 35B-A3B is the first open-source MoE model running losslessly on edge devices using forward-looking expert scheduling. It operates on an iPhone 17 Pro Max with just 3GB memory at 13+ token/s decoding speed, enabling local multimodal reasoning for enhanced privacy and latency. Compatible with iOS, Android, and HarmonyOS across Intel, AMD, Arm, Rockchip, and Huanmo chips, it also offers open-source Self-Replay Fine-Tuning. The 35B-A3B outperforms Qwen3.6 35B-A3B in speed and scores, while the 1B version surpasses Qwen3.5 0.8B, lowering thresholds for large models on smartphones.
0
0
/live/7847
Author DanielPerez DanielPerez October 13, 2026
Apple executed a reverse acqui-hire of AI audio startup Huxe, acquiring its core team and licensing its intellectual property. The move follows Huxe's sudden shutdown after its founders, creators of NotebookLM's AI podcast feature, departed. Apple confirmed the deal to the European Commission on June 9, hiring Huxe employees while securing non-exclusive rights to their tech. This strategic talent acquisition aims to enhance Apple's personalized audio capabilities, potentially integrating Huxe's technology into its Podcasts app following Spotify's recent launch of similar AI features.
0
0
OR