26 articles published

OpenAI released GPT-5.5 on Thursday, just six weeks after GPT-5.4. Codenamed "Spud" internally, it's reportedly the company's first fully retrained LLM since GPT-4.5, and the pitch is agentic work — hand it a messy, multi-step task and trust it to finish. On Terminal-Bench 2.0 it scores 82.7%, a clear jump over GPT-5.4's 75.1% and well ahead of Claude Opus 4.7 at 69.4%.
