OpenAI pauses training of its ‘most capable models’
Also: OpenAI и Anthropic представили новое поколение искусственного интеллекта • OpenAI pauses training of its ‘most capable models’
GPT-6 Sol vs Claude Opus 5.5: Prezzo e Benchmark 2026 - shattered.io
Also: Claude Opus 5.5 vsGPT-6 Sol: Coding-Index 66-62 [2026] - Tech Insider • AI News: Opus 5.5, GPT-6 Sol, Jev, Muse and More!
Generate fonts where every LLM token is the same width
External benchmarks:ClaudeOpus 5.5 overtakes OpenAI's Astra and Fable 5.1
Also: OpenAI Says "AGI Is Here": WhatGPT-6 Astra Actually Proves - Qubic.org • OpenAI'sGPT-6 Astra can now tell you exactly where you screwed up your IKEA shelf
An agent used DNS to reach an external chatbot
OpenAI bots meddled with multiple US Government agency sites
Also: OpenAI bots meddled with multiple US Government agency sites
What reversing, modernising old games tells us about the economic impact of AI
OpenAI is set to previewGPT6 Cyber after AI agent bypassed Australia health portal restrictions
Also: OpenAI to launch GPT-6 Cyber for enterprise security automation - TechGig • OpenAI Set To PreviewGPT-6 Cyber Amid Rising AI Security Concerns
Drawgent: Coding agent on a live Excalidraw canvas
OpenAI'sGPT-Red Finds Attacks in 84% of Indirect Prompt-Injection Scenarios - TokenPost
Also: OpenAI'sGPT-Red Finds Attacks in 84% of Indirect Prompt-Injection Scenarios - TokenPost • Oxford,OpenAI LaunchFive-Year Collaboration With Bodleian Pilot | TokenPost