Flash

The news, the context, and what changes.

RSS
Flash

Meta Superintelligence Lab Launches AI Image Generation Model Muse Image

Meta's Superintelligence Lab has released its first self-developed AI image generation model, Muse Image. The model supports text-to-image creation, editing, and QR code generation, ranking second on the Arena AI leaderboard while providing various free AI effects for Instagram Stories.

88Z AI editorial1 min read
Flash

Ant Lingbo Launches Spatial Perception Model LingBot-Depth 2.0

Ant Lingbo Technology has released the spatial perception model LingBot-Depth 2.0, with training data significantly expanded to 150 million. It performed excellently in depth completion benchmarks. The company also open-sourced the visual foundation model LingBot-Vision and achieved certification cooperation with Orbbec.

88Z AI editorial1 min read
Flash

BAAI Launches Multimodal Representation World Model WuJie·RoboBrain Orca

BAAI has released WuJie·RoboBrain Orca, a multimodal representation world model. The model utilizes 125,000 hours of video for unconscious learning and combines this with 160 million event annotations for conscious learning to construct a unified latent representation space of the world. It outperforms baseline models of similar scale in downstream tasks such as text generation, image prediction, and embodied action generation.

88Z AI editorial1 min read
Flash

Physical AI Company Momenta Officially Lists, Market Cap Exceeds HK$70 Billion

Momenta listed on the HKEX under code 6880.HK. Its opening price rose over 6% on the first day, pushing the market cap above HK$70 billion. The company raised about HK$6.8 billion, with public offering oversubscribed 414 times and participation from 14 cornerstone investors including GIC and Fidelity.

88Z AI editorial1 min read
Flash

ByteDance Launches Ultra-Long-Horizon Benchmark EdgeBench

ByteDance's Seed team has released EdgeBench to evaluate the continuous learning capabilities of AI Agents in real-world environments. The benchmark includes 134 real tasks across six major domains, supporting extended continuous operation. Research indicates that Agent learning trajectories follow a high-precision log-sigmoid curve, with leading models doubling their learning speed every three months.

88Z AI editorial1 min read
Flash

Alipay AI Open Platform Opens for Invited Testing

The Alipay AI Open Platform has officially launched and opened for invited testing, offering a one-stop AI upgrade solution for merchants and developers. Existing applications can be quickly transformed into MCPs, Skills, or Agents, connecting with a billion users via the Abao assistant and covering multiple terminals including phones, car systems, and AI glasses.

88Z AI editorial1 min read
Flash

Tencent's AI App Generator 'Toast' Launches on Apple iOS

Tencent's AI-generated application app 'Toast' has officially launched on Apple iOS, following the Android release last month. Positioned as an exploratory atmospheric programming tool, it enables users without coding skills to turn ideas into functional apps, creating a closed-loop ecosystem from creation to distribution.

88Z AI editorial1 min read
Flash

Shanghai AI Lab Open-Sources Domestic Research Agent Workbench InternAgentS

Shanghai AI Lab has released the open-source research agent workbench InternAgentS, designed specifically for the AI for Science domain. The platform supports local deployment and multi-model adaptation, integrating full-process functions such as paper reading, experiment analysis, code iteration, and scientific writing. It is compatible with domestic and private models like DeepSeek, Qwen, and Kimi, connecting to over 3,600 research tools and skills.

88Z AI editorial1 min read
Flash

Alibaba Launches Qoder Enterprise Edition with Corporate Knowledge Base and Flexible Credits Allocation

Alibaba Cloud has launched Qoder Enterprise Edition, featuring the QMind personal cloud knowledge base for multi-source data integration and cross-dimensional knowledge sharing. The new version introduces a pooled Credits payment model allowing administrators to dynamically allocate quotas, along with five-layer security protection covering transmission encryption to AI runtime, certified under ISO 27001.

88Z AI editorial1 min read
Flash

Meituan Open-Sources LongCat-2.0, Releases Inference Code for Domestic Chips

Meituan has officially open-sourced LongCat-2.0, a trillion-parameter large model designed specifically for Agentic Coding tasks. With 1.6T total parameters and 48B average active parameters, the model introduces innovative sparse attention and N-gram Embedding techniques to enhance long-context processing and code execution performance. It has completed inference validation on a domestic computing cluster of 50,000 cards. The release includes multi-precision versions (BF16/FP8/INT8) and optimiz

88Z AI editorial1 min read
Flash

Tencent Hunyuan Open-Sources Hybrid Expert Model Hy3

Tencent Hunyuan has officially released the Hy3 large model. By improving training data quality and scaling reinforcement learning compute, it achieves significant progress in reasoning, agents, and long-context tasks. Its performance rivals flagship models with 2-5 times more parameters, while substantially reducing hallucination rates, common sense errors, and multi-turn issue rates.

88Z AI editorial1 min read
Flash

Alibaba Open-Sources Real-Time Speech Recognition Model Fun-ASR-Realtime

Alibaba has launched the upgraded Fun-ASR-Realtime, a real-time speech recognition model with sub-second initial latency and accuracy comparable to offline systems. It supports 30 languages and 16 dialects, featuring context awareness and self-correction capabilities, outperforming competitors in dialect recognition tests.

88Z AI editorial1 min read