OpenAI confirms GPT-6 is not shipping in 2025 (www.bleepingcomputer.com)

OpenAI has confirmed that GPT-6 will not ship in 2025, refuting viral claims that the next major model would arrive by year-end. The rumor — sparked by analyst Mark Mahaney and amplified on X — was publicly denied by OpenAI and a pseudonymous employee known as Roon (@tszzl). While the company says GPT-6 won’t debut this year, it hasn’t ruled out continued releases and iterative updates to its current family of models. Technically, OpenAI’s focus remains on evolving the GPT-5 line: GPT-5 Auto dynamically routes queries between “reasoning” models (longer, deeper thinking for complex answers) and faster, non-reasoning variants like GPT-5‑instant, trading latency for response depth as needed. That architecture and the multiple post-launch updates to GPT-5 suggest OpenAI is prioritizing incremental improvements (e.g., potential GPT‑5.5 releases), production robustness, and smarter model orchestration over a sweeping version jump. For the AI/ML community, this tempers expectations around breakthrough leaps and signals continued emphasis on latency/reasoning trade-offs, model routing systems, and iterative performance tuning rather than an imminent GPT‑6 milestone.