Flash

The news, the context, and what changes.

RSS
Flash

Anthropic Launches Strongest AI Models Claude Fable 5 and Mythos 5

Anthropic has released its most powerful model, Claude Fable 5, which outperforms previous generations across the board and leads significantly in benchmarks like SWE-Bench Pro. The model offers higher token efficiency, capable of migrating a 50-million-line codebase in one day. Additionally, Anthropic launched the unguarded version, Claude Mythos 5, available exclusively to Glasswing partners. For safety, Claude Fable 5 includes guardrails that downgrade sensitive requests to Claude Opus 4.8.

88Z AI editorial1 min read
Flash

ByteDance's TRAE SOLO Upgrades to TRAE Work

ByteDance's TRAE announces that TRAE SOLO has officially upgraded to TRAE Work, expanding its brand meaning from The Real AI Engineer to The Real AI Enabler. The new version targets broader real-world work scenarios with Work and Code modes, covers PC, mobile, and Web platforms, supports parallel multitasking, and completed integration with Feishu in May.

88Z AI editorial1 min read
Flash

Ant Group Launches AI App Generation Platform QMuse

Ant Group has launched QMuse, an AI app generation platform that enables zero-code creation of runnable applications via natural language. Currently in beta and requiring an invite code, the platform aims to lower development barriers while ensuring security.

88Z AI editorial1 min read
Flash

Google Launches Real-Time Translation Model Gemini 3.5 Live Translate

Google has released Gemini 3.5 Live Translate, supporting real-time translation for over 70 languages while preserving original tone and speed with only a few seconds of latency. Developers can access the API preview, enterprises can use it in Google Meet, and general users can try it via the Google Translate app.

88Z AI editorial1 min read
Flash

Xiaomi Launches MiMo-V2.5-Pro UltraSpeed Mode

Xiaomi, in collaboration with TileRT, has launched the MiMo-V2.5-Pro UltraSpeed mode, achieving a generation speed of over 1000 tokens/s for trillion-parameter models on general-purpose GPUs. Priced at three times the standard version, it offers approximately tenfold faster output and is available exclusively via API. Access is limited and application-based, prioritizing enterprises and professional developers.

88Z AI editorial1 min read
Flash

AI Short Drama Creator 'Jingying Technology' Announces Completion of Tens of Millions of USD in Series A and A+ Financing

Jingying Technology has completed tens of millions of USD in Series A and A+ financing, with investors including Wang Huiwen's family office and Ant Group. The company also announced that Wang Minjie, former Chief Applied Scientist at AWS, has joined as Chief Scientist. Positioned as the first Agent-native company in the content industry, it focuses on AI short dramas to build a reinforcement learning environment for creator Agents that is accessible and self-evolving, driven by real consumer fe

88Z AI editorial1 min read
Flash

Daxiao Robotics Launches Kairos-HomeWorld, the First Whole-House 3D Interactive World Model

Daxiao Robotics, in collaboration with The Chinese University of Hong Kong and Shenzhen He Tao College, has launched Kairos-HomeWorld, the world's first whole-house 3D interactive world model. Built on a four-stage hierarchical generation architecture, the model generates coherent, physically plausible, and functionally complete 3D home environments from text prompts with a single click. Each scene averages over 15 interactive objects, enabling robots to undergo simulation training for complex h

88Z AI editorial1 min read
Flash

Google and SpaceX Reach $920 Million Cloud Computing Partnership

Google and SpaceX have reached a cloud computing partnership. From October 2026 to June 2029, Google will pay SpaceX $920 million monthly to lease computing power from approximately 110,000 Nvidia GPUs and CPUs for AI training and inference. The agreement aims to ease Google's compute supply constraints while creating a significant revenue stream for SpaceX's AI business, adding new narrative value to its IPO.

88Z AI editorial1 min read
Flash

"Huachao Shenkong" Secures Hundred-Million-Yuan Angel Series Financing

Next-generation AI ultrasound BCI company Huachao Shenkong has completed a hundred-million-yuan angel series financing round, backed by Matrix Partners China, Delian Capital, and Daoyuan Capital. The funds will support the verification and industrialization of its low-intensity focused ultrasound non-invasive brain-computer interface technology, advancing clinical translation for indications such as pain and addiction.

88Z AI editorial1 min read