Independent test validates only one of OpenAI's three GPT-5.5 Instant claims

The New Stack stress-tested OpenAI's three headline claims for GPT-5.5 Instant — the new default ChatGPT model that replaced GPT-5.3 last week. Only one claim fully held up under scrutiny, fueling skepticism about marketing-driven benchmark choices. OpenAI also rolled out product feed ads in ChatGPT for retailers.