On the fourth day of its 28-day consecutive launch or reset campaign on Tibo, OpenAI introduced the Ultrafast mode for GPT-6.1 Sol. This mode approaches the intelligence level of the flagship model Astra, with a generation speed 8 times faster than the standard GPT-6.1 Sol. However, the pricing has been raised accordingly, at $12 per million input tokens and $60 per million output tokens—6 times higher than the standard version, though still far below the estimated price for Astra's Ultrafast mode. Nevertheless, the launch of the Ultrafast mode has sparked widespread dissatisfaction in the community. Users accused OpenAI of deliberately slowing down the base model to indirectly raise prices by selling the high-speed mode. Additionally, the Ultrafast mode is not available to users on the $200 Pro plan, leading to claims that it is irrelevant to most regular users. Previously, GPT-6.1 Sol offered capabilities close to Astra at roughly one-fifth of Astra's standard token price, excelling in programming and computer operation-related evaluations. However, in real-world testing, its generation speed was only 20-30 tokens/s, failing to meet efficiency demands for scenarios like long code generation and multi-step Agent operations. The update also introduced a real-time Codex correction feature, allowing users to intervene and adjust directions during task execution. Combined with the Ultrafast mode, this feature is expected to enhance efficiency in complex workflows. Currently, the Ultrafast mode is only available to users on the Pro 500 subscription plan, eligible enterprise paid plans, educational credit plans, and API users. By separating speed from the default model experience and offering it as a separate paid service, OpenAI's actual value proposition remains to be validated by the market.