OpenAI ships GPT-5.4-Cyber and upgraded Agents SDK

OpenAI

OpenAI released GPT-5.4-Cyber for defensive security work, offering broader access to verified users than Anthropic's competing Mythos model. The system can reverse-engineer software to detect malware and vulnerabilities, representing a more open approach to deploying advanced AI cybersecurity capabilities.

Google Blog

Google introduced 'Skills' in Chrome, allowing users to save and reuse Gemini-powered AI prompts across websites for one-click execution of tasks like summarization and content transformation, deepening AI integration into the browser experience.

Google Blog

Google's Gemini 3.1 Flash TTS delivers significantly improved text-to-speech quality with granular control via audio tags, supporting 70+ languages and featuring SynthID watermarking for AI-generated content detection. The model tops industry benchmarks while offering developers unprecedented creative control over vocal performance and multi-speaker dialogue.

OpenAI

OpenAI's updated Agents SDK provides developers with production-grade infrastructure for building agents, including native sandbox execution, configurable memory, and integrations with major cloud providers. The update addresses the gap between prototype frameworks and production deployment by offering a turnkey yet flexible harness optimized for frontier model capabilities.

Financial Times

Meta is building an AI version of Mark Zuckerberg to interact with employees, showcasing the company's application of its AI technology to create executive digital clones for internal organizational use.