OpenAI Releases GPT-5.5 Instant, Making ChatGPT Faster and with Fewer Hallucinations

OpenAI has officially released a new artificial intelligence (AI) model dubbed GPT-5.5 Instant. This AI model is now the default model in the ChatGPT service. It replaces GPT-5.3 Instant and is focused on everyday usage. In its official blog, OpenAI states that GPT-5.5 Instant is designed to provide faster, more concise answers and to reduce errors or “hallucinations” in AI responses. “The Instant model is now more reliable, with significant improvements in accuracy across various fields,” writes OpenAI. Based on several internal tests, OpenAI claims that GPT-5.5 Instant produces 52.5 per cent fewer hallucination claims compared to GPT-5.3 Instant. The hallucination rate will be lower for high-risk questions such as topics in medicine, law, and finance. GPT-5.5 Instant is also designed to provide more “to the point” answers. OpenAI states that responses from this model use around 30 per cent fewer words and nearly 30 per cent fewer lines, without reducing the informational content. As the default model, GPT-5.5 Instant is used for various general needs, from writing and answering questions to light analysis. Overall, OpenAI states that this model is also better at analysing photos and images, answering science and mathematics questions (STEM), and determining when web searches are needed to enrich answers. In a number of internal benchmarks, OpenAI states that GPT-5.5 Instant shows fairly good performance improvements over GPT-5.3 Instant. However, these results come from internal testing and have not yet been widely compared with other AI models.