Flash

The news, the context, and what changes.

RSS
Flash

OpenAI Opens Free Text Chat with ChatGPT

OpenAI announced that text chats with ChatGPT are no longer restricted for free users, with the default model upgraded to GPT-5.6 Luna, which outperforms GPT-5.5 Instant. Free and Go subscribers can also use a new thinking button for complex problems, but high-compute features like file uploads, image generation, and voice conversations remain limited. Paid users will receive the more reliable and concise GPT-5.6 Sol along with a five-level thinking intensity slider.

88Z AI editorial1 min read
Flash

Volcano Engine Launches Seedance 2.5 API Service

Volcano Engine has officially launched the Seedance 2.5 API, delivering significant improvements over version 2.0 in instruction following, long-form narrative, human-like realism, and audio-visual quality. The service natively supports direct generation of 30-second videos, allows up to 50 full-modality reference materials, and is compatible with over ten languages. The model is now simultaneously available on platforms including LibTV, Qixiang AI, and TapNow.

88Z AI editorial1 min read
Flash

Ant Bailin Launches Native Hybrid Reasoning Model Ling-3.0-tiny

Ant Bailin has released Ling-3.0-tiny, a native hybrid reasoning model with 7.9B total parameters and 1.3B activated per token. Optimized for mathematical reasoning and edge deployment, it supports tool calling and fully local offline inference.

88Z AI editorial1 min read
Flash

Meta Launches On-Device Coding AI Agent Muse Code

Meta has released the beta version of Muse Code, its first coding AI agent tool, led by Chief AI Officer Wang Tao. Built on the Muse Spark 1.2 model, it aims to handle complete software engineering tasks within large codebases, directly challenging Anthropic's Claude Code and OpenAI's Codex.

88Z AI editorial1 min read
Flash

Alibaba Cloud Launches One Key MCP for One-Click Access to Multiple MCP Services

Alibaba Cloud has launched the One Key MCP service, enabling developers to call all ecosystem partners' MCP services with a single Alibaba Cloud Bailian API key. Compatible with mainstream Coding Agents such as Qoder, Codex, Claude Code, and Cursor, it significantly simplifies access, authentication, and billing management for multiple MCP services. Users can log in to the Alibaba Cloud Bailian platform and directly activate and call relevant services in the MCP Square using their existing API k

88Z AI editorial2 min read
Flash

Tencent Launches AI-Native Data Analytics Platform Omega

Tencent has launched Omega, an AI-native data analytics product, with its beta version named 'Marmos'. Users can generate data dashboards via natural language, supporting interactive modifications and dynamic refreshes. The product utilizes QueryRegistry and DTBridge architectures, combined with multi-layer permission verification and AST read-only checks, aiming to resolve data stagnation and security issues in traditional AI dashboards.

88Z AI editorial1 min read
Flash

Alibaba Launches Native Computer Use Agent 'Qwen-CUA'

Alibaba's Qwen team and XLang Lab have jointly released Qwen-CUA, a native computer use agent. Built on a mixture-of-experts architecture, it perceives interfaces via screenshots and directly outputs keyboard and mouse events to control software without relying on underlying metadata. It performs well on the OSWorld-Verified benchmark, with the trillion-parameter version scoring 87.6.

88Z AI editorial1 min read
Flash

Tencent Hunyuan Launches New Speech Recognition Model Hy ASR 3.0 Preview

Tencent Hunyuan has released the new Hy ASR 3.0 preview speech recognition model. Built on the Hy3 large language model, it deeply integrates speech recognition with deep semantic understanding. The model achieves top-tier industry accuracy in Mandarin, English, and Cantonese scenarios, featuring intelligent error correction in complex contexts.

88Z AI editorial1 min read
Flash

Alibaba Qwen Image Generation Model Qwen-Image-3.0 Launches on Qwen AI Platform

Alibaba's Qwen image generation model, Qwen-Image-3.0, has launched on the Qwen AI platform and is freely available to all users. The flagship and standard versions have simultaneously opened their APIs. In the latest text-to-image rankings on the international evaluation platform Arena.ai, the model ranks first in China and fifth globally.

88Z AI editorial1 min read