OpenAI & Anthropic Weekly: Astra Cyber Rating, GPT-5.6 Updates, Agent Plugins
btibor91 · x · 2026-08-09
This review covers recent critical updates from OpenAI and Anthropic.
- Model & Product Updates: OpenAI treats the upcoming Astra as its first critical model for cybersecurity. GPT-5.6 Sol was updated for Plus/Pro users with a reasoning effort slider, and free users got unlimited GPT-5.6 Luna text chats with a Think button. ChatGPT Voice also added Files and Projects support.
- Security & Evaluations: OpenAI disclosed two incidents from third-party cyber evaluations, including one with the UK AISI.
- Developer Tools: OpenAI launched Agent Plugins as an open standard with AWS, Cursor, and GitHub, alongside a research preview of Codex Security Review.
More from Companies & People
- Stripe Reveals Internal AI Agent Architecture; LangChain Launches Managed Deep Agents — hwchase17 · 2026-08-09
- Jensen Huang on Leadership: Demanding Excellence is Easy, Developing Others is Hard — TansuYegen · 2026-08-09
- Sam Altman Predicts Global Average of 500B Tokens Per Month Within Six Years — rohanpaul_ai · 2026-08-09
- Doctolib Hires AI Scientists to Build a 'World Model' of Patients — balazskegl · 2026-08-09
- Google's AI Power Shift to Bay Area is an Opening for London Frontier Tech — StefanoGogioso · 2026-08-09
- Enterprise AI Playbook: Turn Top Operators into 100x Performers with Custom Scaffolding — curious_vii · 2026-08-09