Flash

The news, the context, and what changes.

RSS
Flash

Mango Media and AIsphere Reach Strategic Partnership

Mango Media and AIsphere have signed a strategic cooperation agreement. Leveraging the Mango Panoramic Workstation, they will integrate AIsphere's AI video large model capabilities to promote the application of technology in audio-video creation, AI film-games, and IP development. The two parties will cooperate in content production, interactive entertainment, and content globalization, exploring industrial applications of AI video in short dramas, animation, variety show visuals, and real-time

88Z AI editorial1 min read
Flash

DeepSeek Launches DeepSeek-V4 Preview

DeepSeek has released and open-sourced the preview of DeepSeek-V4, featuring Pro and Flash variants that both support a 1M context window. The Pro version matches top closed-source models in performance, while the Flash version offers similar reasoning capabilities at a lower cost. The new architecture significantly reduces computation and memory overhead for long contexts.

88Z AI editorial1 min read
Flash

OpenAI Launches Latest Flagship Large Model GPT-5.5

OpenAI has released its new flagship model, GPT-5.5, designed for practical work and agents. Co-designed with NVIDIA GB200/GB300 NVL72 systems, it outperforms previous generations and Claude Opus 4.7 in coding, knowledge work, and scientific research. GPT-5.5 breaks performance bottlenecks, matching GPT-5.4's speed while consuming fewer tokens.

88Z AI editorial1 min read
Flash

iFlytek Loomy Fully Opens, Enabling 'Plug-in' Office Mode

iFlytek's AI office assistant Loomy is now fully open for registration without invitation codes. Having completed over 5,000 task loops across more than 20 industries including finance and e-commerce, the latest update introduces natural language control for IM tools, guided PPT creation, and parallel task processing to boost efficiency.

88Z AI editorial1 min read
Flash

StepFun Launches New Automatic Speech Recognition Model StepAudio 2.5 ASR

StepFun has released StepAudio 2.5 ASR, incorporating large language model inference acceleration technology. Built on the ASR+MTP-5 architecture, the model significantly improves speed and reduces costs, achieving SOTA levels in Chinese and English evaluations while supporting transcription of 30-minute long audio clips.

88Z AI editorial1 min read
Flash

Alibaba Cloud Launches Enterprise Agent Building Platform JVS Crew

Alibaba Cloud has released JVS Crew, a platform designed to help enterprises embed production-grade AI Agent capabilities into existing applications through an 'integrated' approach. The platform strengthens security and control in enterprise scenarios by providing triple protection for identity, content, and execution, along with managed memory, end-to-end observability, and multi-level cost control.

88Z AI editorial1 min read
Flash

OpenAI Launches Team Collaboration Agents: Workspace Agents

OpenAI has introduced Workspace Agents in ChatGPT, enabling teams to create collaborative agents for complex tasks. Powered by Codex, these agents handle files, run code, and store memory, operating 24/7 in the cloud with Slack integration. The feature is currently available for Business, Enterprise, Edu, and Teachers plans.

88Z AI editorial1 min read
Flash

Alibaba Tongyi Open-Sources Dense Multimodal Model Qwen3.6-27B

Alibaba's Tongyi Qianwen team has open-sourced Qwen3.6-27B, a dense multimodal model with 27 billion parameters. Supporting both multimodal thinking and non-thinking modes, it achieves flagship-level breakthroughs in agent coding capabilities, outperforming the previous open-source flagship Qwen3.5-397B-A17B on benchmarks like SWE-bench and Terminal-Bench. Its dense architecture allows deployment without MoE routing, facilitating broader adoption.

88Z AI editorial1 min read
Flash

ByteDance Launches New 3D Generation Large Model Seed3D 2.0

ByteDance has released Seed3D 2.0, achieving state-of-the-art performance in geometry and texture metrics. The model employs a two-stage DiT strategy and a unified PBR architecture, integrating MoE and VLM priors to enhance material realism. It also expands capabilities such as part decomposition, moving 3D generation toward production readiness.

88Z AI editorial1 min read
Flash

Tencent Hunyuan Open-Sources Hy3 Preview, a Mixture-of-Experts Model Integrating Fast and Slow Thinking

The Tencent Hunyuan team has released the Hy3 preview model, marking their first achievement after rebuilding training infrastructure. Featuring a Mixture-of-Experts architecture that integrates fast and slow thinking, the model has 295B total parameters, 21B activated parameters, and supports a 256K context window. As the most intelligent model in the Hunyuan series, it excels in challenging reasoning benchmarks, with significantly enhanced agent and coding capabilities, enabling the one-shot g

88Z AI editorial1 min read
Flash

StepFun and Qianli Technology Announce Comprehensive Strategic Partnership

StepFun and Qianli Technology have formed a comprehensive strategic partnership to co-develop a native intelligent driving foundation model. StepFun's open-source flagship model, Step 3.5 Flash, has achieved leading results in various benchmarks, backed by its deep layout in Agent foundation models and omni-modal technologies.

88Z AI editorial1 min read