OPENAI PULLS THE PLUG ON GPT-4.5 — MODEL GOES DARK JUNE 27 AS THE COMPANY RACES TO PUSH OUT 5.6 OpenAI confirmed this week that GPT-4.5 will be retired from ChatGPT and the API on June 27, 2026, following a 30-day sunset period. The model, which was positioned as a bridge between GPT-4 and the GPT-5 generation, will stop accepting new requests at that date and any integrations that have not migrated will break. The timing is deliberate. GPT-5.6 is targeting a June launch, bringing what OpenAI describes as a significant step forward in reasoning capability and autonomous agent workflows. The company is building toward a model that requires less human oversight for repetitive decision chains, a priority that reflects where enterprise customers are pushing the hardest. GPT-5.5 Instant is the current default in ChatGPT. OpenAI overhauled it earlier this month, describing the updated version as more natural in everyday conversation and better calibrated for practical tasks, with fewer bullet-heavy or overly long responses. That update addressed a persistent complaint from users who felt the model was performing for an audience rather than talking to a person. For developers, the deadline is real. GPT-4.5 had distinct output characteristics that some applications depended on, and GPT-5.5 Instant behaves differently. Anyone running production workloads on the older model has three weeks to migrate or accept the consequences. Keywords: GPT-4.5 retirement, OpenAI GPT-5.6, ChatGPT update, OpenAI model deprecation