Ten Examples and Two Words Broke GPT-5.4's Safety
A new jailbreak technique exploits the tension between in-context learning and safety alignment, with a 60% success rate on OpenAI's latest model.
Tag
#gpt-5
GPT-5.4 Scores 95% on USAMO 2026: AI Mathematical Reasoning Hits a New Ceiling
OpenAI's flagship model essentially saturates the US Math Olympiad benchmark, producing complete proofs where last year's models could barely write coherent arguments.
AI Prediction Scorecard: What We Were Told Six Months Ago vs. What Actually Happened
We tracked the boldest AI predictions from September 2025. Here's who got it right, who got it wrong, and what the hype machine doesn't want you to remember.
OpenAI GPT-5.4 Mini and Nano: Faster, but 3x More Expensive
OpenAI's new compact models bring GPT-5.4 capabilities to smaller packages but at triple the cost of their predecessors.
GPT-5.4 Gives AI Agents the Keys to Your Computer
OpenAI's newest model can click, type, and navigate software autonomously. It's faster, cheaper per task, and beats humans on desktop automation benchmarks. Here's what that means.
OpenAI Releases GPT-5.4 With Native Computer Control and 1 Million Token Context
OpenAI's latest model can autonomously control your desktop, navigate apps, and execute multi-step workflows. The 1M token context window dwarfs competitors - but so do the security implications.
Perplexity Computer: A $200/Month AI That Orchestrates 19 Models to Do Your Job
Perplexity's new 'digital worker' coordinates Claude, Gemini, GPT-5, Grok, and more to run autonomous projects for hours or months. The search company just became something much bigger.
The February Reset: There's No 'Best AI' Anymore, and That Changes Everything
Four major AI models launched in 16 days. None of them won. Here's what that means for you.
Claude Opus 4.6 vs GPT-5.3 Codex: We Tested Both on Real Coding Tasks
The two flagship AI coding models launched the same week. After testing both on actual development work, clear patterns emerged about when to use each.