Breaking

OpenAI Adds Astra Ultrafast Mode to Codex for Faster AI Coding Sessions

September 29, 2026 at 14:07 EDT

Software Dev & Coding
AI Agents
Foundation Models

OpenAI has introduced Astra Ultrafast for Codex, a premium speed setting that can generate GPT-6 Astra tokens up to 8x faster than Astra Standard and up to 4x faster than Astra Fast. The move extends OpenAI's fastest Astra experience from the API into its software-development agent, targeting developers who use Codex in tight edit, test and review loops.

Sep. 29, 2026 · OpenAI

Astra gets an Ultrafast lane

Codex can now run GPT-6 Astra in a premium mode built for faster interactive coding. The headline claim is token generation up to 8x faster than Standard, with higher usage and billing multipliers.

Speed modes, drawn to scale

Standard

Fast

Ultrafast

The comparison is token generation speed, not full task completion time.

vs Standard speed

vs Fast speed

included usage rate

paid credit rate

Where it fits

Best use

Interactive coding moments where waiting blocks iteration.

Access

Pro $500 plus eligible Enterprise and Edu plans.

Limit

Not for workspaces requiring inference residency outside the United States.

Upside

Faster token flow can make Codex feel more responsive during edit, test and review loops.

Tradeoff

Usage and paid-credit multipliers rise, so teams still need to reserve it for high-value moments.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.