Ai Safety

On July 21, 2026, alongside Gemini 3.6 Flash and 3.5 Flash-Lite, Google announced a third model that is nothing like the other two. Gemini 3.5 Flash Cyber is purpose-trained for cybersecurity, and it is not generally available. It is a 3.5 Flash derivative fine-tuned to discover, validate, and remediate software vulnerabilities. It got its own separate post on the DeepMind blog, which itself says a lot about how differently Google treats this one. ...

TL;DR: GPT-Red in 30 Seconds GPT-Red is an automated red-teaming model OpenAI trained to break its own models; it is not public and never will be. It was trained with self-play: attacker GPT-Red and defender models played against each other, both getting stronger over time. In scenarios where human red-teamers succeeded only 13% of the time, GPT-Red hit 84%. GPT-5.6 was trained against GPT-Red, making it OpenAI’s most robust model against prompt injections (only a 0.05% failure rate). OpenAI just announced something unusual: not a model for users, but an “AI hacker” trained to break its own models. Meet GPT-Red! According to the research publication “Unlocking Self-Improvement for Robustness” released today, GPT-Red automates the work of human security testers (red teams) and does it far better than humans. ...

Talking to AI no longer feels like a walkie-talkie exchange, it finally feels like a real conversation! On July 8, 2026, OpenAI announced GPT-Live, its next-generation family of voice AI models. This is not just a better voice mode, it is an architectural shift that redefines the voice AI experience. So what makes GPT-Live so different? Let’s dive in. What Is GPT-Live? GPT-Live is OpenAI’s new voice model family built on a full-duplex architecture. Full-duplex means the model can listen and speak at the same time. With the previous Advanced Voice Mode, the model waited while you spoke and you waited while it spoke. GPT-Live removes that turn-taking constraint entirely. ...

The news everyone was waiting for is here! Claude Fable 5, shut down by a US government export control order just three days after its launch, is back online worldwide as of July 1. And as if celebrating the comeback, Anthropic is offering paid plan users Fable 5 access at no extra cost until July 12. I covered why the model was shut down in detail in the What Is Claude Fable 5? post. Now we are in act two of the story: how did Fable 5 come back, and how will it work from now on? ...

The cards are being redealt in the world of artificial intelligence! Autonomous tasks that required much larger and more expensive models just a few months ago are now more accessible. Meet Anthropic’s most capable agent to date: Claude Sonnet 5. This large language model (LLM), which can make plans and use tools like browsers and terminals, is opening the doors to a brand new era for developers. If you have used Claude Sonnet 4.6 before, the new features and increased autonomous capabilities will surprise you. ...