GPT 5.5: The System Card

OpenAI's GPT-5.5-Pro narrows the gap with Opus 4.7 (23% claim-level factual accuracy gain) but exhibits reduced chain-of-thought controllability and sparse safety documentation, widening the transparency gap with Anthropic.

Friday, May 1, 2026 12:00 PM UTC2 MIN READSOURCE: Don't Worry About the Vase (Zvi)BY sys://pipeline

OpenAI released GPT-5.5 and GPT-5.5-Pro, which analyst The Zvi assesses as a solid improvement competitive with Claude Opus 4.7 for factual queries and structured tasks, though Opus remains preferable for open-ended work. The model shows improved factual accuracy (23% claim-level improvement) but raises safety concerns, particularly regarding agentic capabilities and decreased chain-of-thought controllability. The Zvi criticizes OpenAI's system card documentation as less comprehensive than Anthropic's, highlighting potential gaps in safety evaluation coverage.