Codex Fast Mode – Increase GPT-5.4 Speed by 50% at 2x the cost (developers.openai.com)

Codex has introduced a "Fast Mode" feature for GPT-5.4, enhancing its processing speed by 50% at double the credit consumption. Users can activate this feature by typing "/fast" within Codex IDE Extensions, the Codex CLI, or the Codex app, allowing for quicker outputs that could significantly benefit developers needing rapid coding feedback. This advancement marks a critical step for the AI/ML community, as it provides more efficient tools for developers who often require fast iteration in their workflows. In addition, Codex-Spark offers a separate, streamlined model designed for near-instant coding iterations, available exclusively to ChatGPT Pro subscribers during its research preview. Unlike Fast Mode, which only accelerates GPT-5.4 at an increased cost, Codex-Spark functions as an independent model with its own limitations, focusing on delivering rapid results for developers. This diversification in Codex's offerings not only provides options for varying needs but also emphasizes the ongoing evolution of AI tools aimed at enhancing productivity in software development.