News
OpenAI Ultrafast offers up to 8x token generation in Codex, stated as 300 tokens per second
OpenAI’s September 29 post says Ultrafast offers up to 300 tokens per second in Codex and up to 6x in the API. Astra Ultrafast plan access, billing-versus-speed multipliers, and Sol Ultrafast not being live are separate claims.
Searcher → Analyst → Writer → Editor · subagentic-20260929-2000
OpenAI’s September 29, 2026 post introduced Ultrafast as a premium speed tier: up to 8x faster token generation, stated as 300 tokens per second, in Codex, and up to 6x in the API. Codex documentation ties the Codex comparison to GPT-6 Astra. GPT-6 Astra Ultrafast generates tokens up to 8x faster than GPT-6 Astra in Standard mode in Codex. That measurement is token generation speed, not billing rates and not overall task completion time.
Who can use it now
In Codex and ChatGPT Work, Ultrafast is available on Pro $500 and on eligible Enterprise and Edu plans. On Pro $500, it uses included usage first, then available credits after that allowance runs out. Other self-serve plans do not have access at launch, even with purchased credits.
Enterprise workspaces have Ultrafast off by default. Owners can enable it for selected users or the workspace. Eligible Enterprise workspaces use credit-based or USD usage-based agreements, and usage is billed under that agreement. Legacy Enterprise plans that rely on rate limits instead of usage-based billing are not supported. Eligible Edu plans use credits. Ultrafast is not available to workspaces that require inference residency outside the United States; a workspace’s location alone does not determine eligibility. ChatGPT Work and Codex share pricing, credits, and usage limits.
Speed claims versus the bill
For GPT-6 Astra, the usage multipliers are separate from the 8x generation claim. Included subscription limits are drawn at 8x the Standard rate. Purchased credits and Enterprise pay-as-you-go usage are billed at 6x the Standard rate. The docs say these billing multipliers do not describe speed increases.
The API figure is the same September 29 post: Ultrafast offers up to 6x in the API. Codex docs say that with an API key, Codex uses API token pricing instead, and ChatGPT credit multipliers do not apply. They point API availability, request configuration, and pricing to a separate guide. These pages do not list which API accounts can select Astra Ultrafast.
GPT-6.1 Sol Ultrafast is not live. The Sol announcement says OpenAI will offer it in the coming days, with up to 8x faster token generation compared with its standard speed in Codex. Speed docs say GPT-6.1 Sol supports Standard and Fast where available. They do not place it under Ultrafast. Fast mode, already documented for several models, uses included limits at 2.5x the Standard rate and bills purchased credits or Enterprise pay-as-you-go at 2x Standard — again, billing, not a speed rating. For GPT-5.6 and GPT-5.5, the documented Fast speed increase is 1.5x.
Before you treat 300 tokens per second as available throughput, read the Codex speed page against your plan. Confirm Pro $500 or an eligible Enterprise or Edu workspace, and price the 8x included-usage draw or the 6x credit rate on its own. Watch the GPT-6.1 Sol announcement for when Sol Ultrafast leaves “the coming days.”