Flash

The news, the context, and what changes.

RSS
Flash

Tencent Hunyuan Launches New Flagship Model Hy4 Preview

Tencent Hunyuan has released its new flagship model, Hy4 preview, featuring 770B total parameters and 49B activated parameters with a 1M context window. The model excels in coding, office tasks, and scientific research, boosting inference throughput by 31.8% and achieving breakthroughs in molecular dynamics simulation speed and geometric problem-solving.

88Z AI editorial1 min read
Flash

Google Launches AI Video Generation Model Gemini Omni 1.1 Flash

Google has released the Gemini Omni 1.1 Flash video generation model, supporting output resolutions up to 4K. The model allows users to extend existing videos in 10-second increments, with a maximum single video length of 40 seconds, and enables smooth transitions by specifying start and end frames. It offers a 60% increase in generation speed at one-third the cost, supports multi-tier output from 360p previews to 4K ultra-high definition, and can incorporate reference videos up to 3 seconds lon

88Z AI editorial2 min read
Flash

Zhipu Open-Sources Native Multimodal Model GLM-5.3-Flash

Zhipu has released GLM-5.3-Flash, the first native multimodal model in the GLM-5 series. It matches the intelligence level of Claude Opus 4.8 at a fraction of the cost. The model uses a hybrid architecture to significantly reduce long-context processing overhead and natively supports vision, code, and GUI collaboration.

88Z AI editorial1 min read
Flash

Alibaba Tongyi Launches Multimodal MoE Model Qwen3.8-Flash

Alibaba Tongyi has released Qwen3.8-Flash, a multimodal MoE model with 125B total parameters and 6B activated per token, supporting ultra-long contexts. With four major technical upgrades, training costs are significantly reduced, coding and office capabilities are enhanced, and competitive API pricing is offered.

88Z AI editorial1 min read
Flash

Tencent Hunyuan Launches On-Device Translation Model Hy-MT2-1.8B

Tencent Hunyuan has released the on-device translation model Hy-MT2-1.8B, which uses proprietary compression technology to drastically reduce model size while maintaining high-quality translation that outperforms several commercial APIs. The model has been optimized for x86 operators and successfully deployed for real-time danmaku translation on Bilibili live streams, supporting 33 languages with a processing time of only 500-800 milliseconds per item.

88Z AI editorial1 min read
Flash

Alibaba Qwen Office International Edition Opens Public Beta

QwenWork International Edition has launched its public beta, offering web and PC access to overseas individual and enterprise users. Positioned as an all-in-one AI productivity platform, it integrates with tools like Slack and Notion, supports multiple languages and account registration methods, and plans to deepen connections with enterprise databases and workflows.

88Z AI editorial1 min read
Flash

Google Launches Speech-to-Text Model Gemini 3.5 Transcribe

Google releases Gemini 3.5 Transcribe, which automatically removes filler words and self-corrections. It is about 70% faster than Chirp 3, with a real-time error rate of 5.5%. The model supports 85 languages and multi-speaker audio, and is already used in Pixel 11 Gboard.

88Z AI editorial1 min read
Flash

PixVerse R2: A Real-Time Omni-Modal World Model by PixVerse

PixVerse has released the technical report for PixVerse R2, building on R1 to further explore scaling paths for real-time world models. Using the Omni Causal AR unified architecture, it supports multi-modal inputs including text, reference images, audio, and motion, enabling continuous world operation with simultaneous generation, interaction, and evolution.

88Z AI editorial1 min read
Flash

Stability AI Completes $76 Million Series B Funding

Stability AI has announced the completion of a $76 million Series B funding round, with proceeds allocated to creative product development and research. Investors include EA, Sony Music, Universal Music, Warner Music, and AMD's venture capital arm.

88Z AI editorial1 min read
Flash

Google Launches Real-Time Collaborative Vibe Coding Tool Play with Putty

Google Labs has released Play with Putty, a real-time collaborative tool focused on the Vibe Coding experience. Users without programming skills can describe requirements in natural language, allowing AI to generate code and render it into functional tools or websites in real time. The tool supports multi-user collaboration via shared links with instant synchronization of changes, aiming to transform software development from individual coding into collective dialogue for rapid prototyping.

88Z AI editorial1 min read