Welcome to the "AI Daily" section! This is your guide to exploring the world of artificial intelligence every day. Every day, we present you with the latest content in the AI field, focusing on developers to help you understand technical trends and innovative AI product applications.
Fresh AI products Click to learn more:https://app.aibase.com/zh
1. Doubao is expected to launch a paid version in mid-June: Three subscription price tiers exposed, will collaborate with e-commerce in the third quarter for traffic diversion
Doubao plans to officially launch paid content in mid-June and update features simultaneously at the Force conference, marking a substantive progress in commercialization for the largest domestic large model. In the third quarter, Doubao will refine paid scenarios by integrating e-commerce functions and drive traffic to Douyin Mall through targeted subsidies. By the fourth quarter, it will fully enter a stable operation phase. Additionally, Doubao will not set the paid user penetration rate as a core performance metric in 2026.
【AiBase Highlights:】
📅 Doubao plans to launch paid content in mid-June and update features simultaneously.
🛒 In the third quarter, it will refine paid scenarios by integrating e-commerce functions and drive traffic to Douyin Mall through targeted subsidies.
📈 Doubao will not set the paid user penetration rate as a core performance metric in 2026.
2. Kuaizi 3.0 officially launched, supporting multi-person and multi-Agent collaboration
The newly launched Kuaizi 3.0 version from ByteDance has achieved significant breakthroughs in team collaboration and ecosystem integration, supporting flexible Agent combination modes, enhancing cross-device synchronization and local ecosystem integration, and providing customized solutions for multiple vertical industries.
【AiBase Highlights:】
🧠 The new version supports flexible combination modes such as "one person + multiple Agents" and "multiple people + multiple Agents".
🔄 Enables seamless switching with mainstream environments and supports various local Agent integrations.
💼 Provides premium templates for core scenarios in multiple industries to help users quickly build expert-level intelligent agents.
3. Krea 2 LoRA now fully open! The era of personalized AI image creation has arrived
The full release of Krea 2 LoRA marks the arrival of the era of personalized AI image creation, allowing users to train and use custom style, character, or object LoRA without a high-end subscription, greatly lowering the barrier to AI image creation.
【AiBase Highlights:】
🎨 Krea2LoRA training function is now fully open, allowing users to customize style, character, or object LoRA.
⚡ Krea2 model centers around a high-performance image generation engine, supporting quick response to vague prompts and generating high-quality images.
🚀 Krea2LoRA significantly outperforms other models in precision, quality, and creative expression, enhancing the personalization capability of AI image creation.
4. OpenAI makes major adjustments to ChatGPT, supporting early locking of computing power levels
OpenAI has made systematic upgrades to ChatGPT, including gesture interaction logic, long conversation directory functions, and underlying model optimization, improving user experience and productivity tool efficiency.
【AiBase Highlights:】
Introducing advanced gestures, allowing users to select different response levels to suit different task needs.
Introducing long conversation directory navigation, enabling users to quickly jump to specific discussion topics.
Optimizing GPT-5.5Instant model, improving text tone and readability, reducing mechanical responses.
5. Programming and GUI both supported! Qwen3.7-Plus launches, completing real app development in 11 hours autonomously
Alibaba officially released the new multimodal large model Qwen3.7-Plus, which has comprehensively upgraded its visual-language capabilities based on text abilities and unified them into an integrated intelligent agent base. Qwen3.7-Plus can seamlessly integrate GUI and CLI interactions, achieving end-to-end automation from front-end prototypes to complex software engineering. In the authoritative visual model ranking Vision Arena, Alibaba has successfully entered the top five globally and first in China.
【AiBase Highlights:】
🧠 Pure text and reasoning Agent: Strong performance in complex software engineering and scientific programming tasks.
🖼️ Multimodal reasoning and visual programming: Supports converting images, videos, and UI screenshots into executable code with one click.
📹 Real-world perception and video understanding: Includes document parsing, advanced OCR, and understanding of long video event streams.
6. WeChat AI agent exposed: Focuses on in-app task processing, expected to be rolled out in phases
Tencent is testing a prototype AI agent product aimed at helping users efficiently complete multiple tasks within WeChat. The launch of this agent will activate the ecological potential of WeChat as a super app and may reshape the service connection mechanism within WeChat.
【AiBase Highlights:】
📱 WeChat is testing an AI agent prototype to improve user task processing efficiency.
🔄 This agent will be rolled out in stages, starting with closed testing and gradually expanding the scope.
🌐 Tencent's move will fully activate the ecological potential of WeChat and promote mobile internet to transition toward AI Agent as the entry point.
7. Disrupting traditional interaction! OpenAI demonstrates "No-App" phone on-site, all interfaces generated in real-time by AI
The article introduces OpenAI's "Agentic operating system" prototype demonstrated at the event, which generates interfaces in real-time through voice commands, eliminating the need to install third-party apps. Heavy inference tasks are handled by cloud-based GPT large models, showcasing a disruptive transformation in future smartphone interaction methods.
【AiBase Highlights:】
📱 The system generates interfaces in real-time through voice commands, without the need to install any third-party apps.
☁️ Heavy inference tasks are processed by cloud-based GPT large models for efficient operations.
📅 OpenAI plans to mass-produce AI agent phones in the first half of 2027, accelerating industrialization.
8. GitHub Copilot officially launches pay-per-use billing mode
GitHub Copilot officially launched the pay-per-use billing mode, introducing an AI point deduction mechanism, marking its transformation from a single code completion tool to an intelligent agent platform.
【AiBase Highlights:】
💡 GitHub Copilot now uses a pay-per-use model, where users must consume AI points to pay for services.
🔄 This model changes the previous low-cost unlimited monthly subscription approach, shifting costs to the user side.
💰 High-value tasks such as code reviews and cloud-based intelligent agent tasks will count toward point consumption, affecting the budget of heavy users.
