OpenAI enhances ChatGPT's health capabilities with GPT-5.5 Instant

More than 230 million people use ChatGPT weekly to ask health and wellness questions, such as understanding lab results or preparing for appointments. [1] The release of GPT-5.5 Instant improves the model's ability to recognize when urgent care is needed, ask for relevant context, and explain uncertainty. [1] Because the model is available to all free ChatGPT users, more individuals can access these capabilities. [1]

OpenAI measures the model's progress using health-specific assessments like HealthBench and HealthBench Professional. [1] These evaluations rely on realistic health conversations and physician-written rubrics to grade accuracy, safety, and communication. [1] On an aggregate of these health evaluations, GPT-5.5 Instant performs at a level similar to OpenAI's frontier Thinking models. [1] The May 2026 release of GPT-5.5 Instant marks a substantial improvement from the previous GPT-5.3 Instant model released in March 2026. [1]

OpenAI is systematically optimizing its models for high-stakes domains by involving physicians in the evaluation loop. By pushing these capabilities to the free GPT-5.5 Instant tier rather than restricting them to premium subscriptions, OpenAI ensures widespread access to clinical-grade reasoning for everyday queries. The model's parity with OpenAI's frontier Thinking models on benchmarks like HealthBench Professional suggests a strategic focus on making efficient models perform exceptionally well on specialized tasks, showing rapid iteration since the GPT-5.3 Instant release in March 2026. What the sources don't address: Whether this model improvement alters OpenAI's liability posture or official terms of service regarding medical advice.