OpenAI Launches Codex Updates with Decisions API and Ultrafast Tier at DevDay

OpenAI introduced significant updates to Codex during DevDay 2026, including a new Decisions API that operates ten times faster than GPT-6 Luna. The Ultrafast tier, powered by GPT-6 Astra, accelerates token generation to 300 tokens per second and is priced at $60 per million input tokens and $300 per million output tokens, six times the standard rate.

The Pro 500 subscription, which costs $500 per month, offers enhanced capabilities, including a limit of 100 messages per week for GPT-6 Pro in chat. Existing Pro plan users will now experience only ten times the usage of the Plus plan in ChatGPT Work and Codex.

OpenAI's Codex now supports operations across devices, allowing users to work from their computers, phones, or the cloud. The Codex Security Cloud feature tracks vulnerabilities from discovery to resolution, enhancing cybersecurity measures. This follows the July 2026 release of the Codex Security CLI as an open-source tool.

OpenAI also announced a partnership with Amazon for an agent platform, while TypeSafe AI introduced the Jev model, which focuses on fast decision-making. The API is currently in limited preview, with broader availability expected shortly.