OpenAI releases GPT-5 and begins phasing out older models

OpenAI has launched GPT-5, marking a shift from a standard language model to an autonomous agent capable of managing entire workflows. In 2026, the technology features deep integration with operating systems, moving toward an ‘AI-first’ ecosystem where voice and context serve as primary commands. A significant advancement is the implementation of hybrid local processing, allowing much of the intelligence to reside on the user's device. This capability addresses European Union privacy requirements by ensuring sensitive data can remain within national territories.

Concurrently, OpenAI has begun decommissioning several older models via its API. As of October 23, 2026, specific snapshots including GPT-4-0613, GPT-4-Turbo, and certain o1 versions have been retired. Developers using hardcoded identifiers must update their integrations to avoid service interruptions. A further round of deactivations for specific GPT-4o and GPT-4.1 snapshots is scheduled for January 14, 2027.