Flash

The news, the context, and what changes.

RSS
Flash

Ant Group Open-Sources Image Generation Model Ming-Image-0.1-Design

Ant Group's inclusionAI has open-sourced the Ming-Image-0.1-Design series, featuring two 6-billion parameter models. The Design model generates complete UI and poster designs from text prompts, while the Layer model splits designs into 2 to 9 independently editable transparent layers. The series achieved top rankings in authoritative benchmarks and demonstrated significantly faster inference speeds compared to larger models.

88Z AI editorial1 min read
Flash

China Telecom Xingchen Lab Open-Sources Document Parsing Model TeleOCR

China Telecom Xingchen Lab has open-sourced TeleOCR, a document parsing model using a lightweight vision-language architecture with approximately 1.2 billion parameters. It uniformly processes digital and photographed documents, converting text, layout, tables, formulas, and scientific charts into structured results. The model topped the OmniDocBench v1.6 leaderboard with a total score of 96.87, surpassing competitors like MinerU2.5-Pro, and also secured first place on PureDocBench and won the I

88Z AI editorial1 min read
Flash

Tencent QClaw WeChat Remote Work AI Assistant to Cease Operations

Tencent announced that due to business adjustments, the QClaw WeChat remote work AI assistant will officially cease operations on December 24, 2026. As of now, new user registrations and subscription renewals have been suspended. Existing users' service benefits remain valid until expiration, and refunds can be requested for unused points. The official recommendation is for users to back up data locally or migrate it to the WorkBuddy platform via the official website.

88Z AI editorial1 min read
Flash

OpenAI Launches Two Models: GPT-6 Sol & GPT-6 Luna

OpenAI has released the GPT-6 Sol and Luna models, cutting API prices by half compared to the 5.6 series. Luna's input price is lower than DeepSeek's off-peak rate, achieving GPT-5.6 Sol performance at about one-tenth the cost. Sol approaches Astra's performance with improved caching and inference efficiency, at a much lower cost than Claude Opus 5 and Fable 5.1.

88Z AI editorial1 min read
Flash

Anthropic Launches Latest Flagship Model Claude Opus 5.5

Anthropic has released Claude Opus 5.5, the first model in its Claude 5.5 series, setting new state-of-the-art records in agent coding, computer use, and knowledge work. It scored 66.4% on Terminal-Bench 4.0, outperforming GPT-6 Astra by 8.5 percentage points, and achieved 1846 Elo on the GDPval benchmark. Additionally, the model is 40% cheaper, 30% faster, and features significantly improved writing and communication capabilities.

88Z AI editorial1 min read
Flash

PixVerse R2: Aishix Technology Launches General Real-Time World Model

Aishix Technology has launched PixVerse R2, the world's first general real-time world model with scaling capabilities. It supports applications in gaming, interactive film, and digital humans. By leveraging Omni Causal AR expansion and real-time acceleration, the model balances real-time performance with generality while enhancing quality, long-range consistency, and multimodal control.

88Z AI editorial1 min read
Flash

Alibaba Launches Qwen Intelligence Full-Stack Solution for AI Phones

Alibaba has released Qwen Intelligence, a non-hardware full-stack AI solution for phone manufacturers. It includes scenario-optimized models, a customizable platform, and three core agents, significantly improving task execution efficiency and generation speed.

88Z AI editorial1 min read
Flash

AiPy Officially Listed on KylinOS Software Store

AiPy has officially launched on the KylinOS Software Store with one-click installation. Combining large models with a Python environment, it executes tasks like data analysis and document generation via voice or natural language commands. It supports multiple platforms, local intranet deployment, and domestic large models, featuring enterprise-grade security capabilities.

88Z AI editorial1 min read
Flash

Kimi Launches Kimi Code Desktop Client

Moonshot AI has officially launched the Kimi Code desktop client, with simultaneous support for macOS and Windows. The application brings AI agent capabilities to a graphical interface, enabling developers to handle requirement expression, progress tracking, debugging, and Git collaboration within a single window. This reduces the barrier to entry and suits various scenarios ranging from prototype development to complex engineering refactoring.

88Z AI editorial1 min read
Flash

SpaceXAI Launches Latest Flagship Large Model Grok 4.7

SpaceXAI has officially released its flagship model, Grok 4.7, featuring 2.1 trillion parameters, a roughly 40% increase from the previous generation. The model shows enhanced capabilities in programming and knowledge work, outperforming GPT-5.6 Sol Max in several benchmarks. However, it trails behind Claude Fable 5.1 and OpenAI GPT-6 Astra. While lagging in specific tests like Terminal-Bench, it ranks first in the second tier for comprehensive intelligence index.

88Z AI editorial1 min read
Flash

Xiaomi Open Sources Full-Modal Model Series Xiaomi MiMo-V2.6

Xiaomi has released and open-sourced the Xiaomi MiMo-V2.6 series, featuring two native full-modal models: Pro and Flash. After extensive reinforcement learning training, the Pro version surpasses several well-known models in comprehensive intelligence index, becoming the strongest open-source model, though it still lags behind top closed-source counterparts.

88Z AI editorial1 min read