OpenAI introduces Ultrafast: up to 8x faster inference, plus a $500 Pro plan

#What happened

What it is: Ultrafast is a premium speed tier, not a separate model. OpenAI demonstrated it at DevDay 2026 with GPT-6 Astra, running matching prompts to build and launch a small rocket scene side by side — the Ultrafast version finished visibly earlier.

The numbers: OpenAI advertises up to eight times faster token generation in Codex, reaching 300 tokens per second, and up to six times faster through the API. This measures token generation speed only — it does not mean an entire job with browsing, compilation and tests finishes eight times faster. GPT-6.1 Sol support is listed as coming soon.

The price of speed: Premium API pricing applies — OpenAI's price list puts Astra Ultrafast at $60 per million input tokens and $300 per million output tokens, six times Astra's standard $10 and $50.

Pro 500: In ChatGPT, Ultrafast is bundled into the new Pro 500 plan at $500 per month, described as 25 times the usage of ChatGPT Plus. Plan tiers are Pro 100 ($100), Pro 200 ($200) and Pro 500 ($500); only Pro 500 includes Ultrafast. OpenAI also reopened Pro 200 sign-ups, which it had paused in September — but new Pro 200 subscriptions get a lower included usage allowance than before, and existing subscribers keep their old allowance until October 29, 2026.

#Why it matters

Sources: https://superpowerdaily.com/posts/everything-openai-announced-devday-2026-keynote and https://madrobot.blog/2026/09/29/openai-devday-2026-everything-announced-dots-gpt-6-1-sol-pro-500/ — Verified September 29, 2026.