According to CoinMeta, on the first day of its 28-day sprint, OpenAI announced that the default generation speeds for GPT-6, Astra, GPT-6.1, and SOL had been optimized, with an overall improvement of about 50%. This speed increase not only applies to Codex and ChatGPT Work, but it will also be automatically reflected in third-party products that use ChatGPT subscription quotas, such as Opencode, PI, AMP, and Devin. Users do not need to make any adjustments; the update will be pushed to them automatically. OpenAI previously promised to deliver an update that would significantly improve the experience for most Codex / Work users every day for the next 28 days, or to reset their quotas directly. This speed increase marks the first day of this sprint.