@ryan_nookpi/pi-extension-codex-fast-mode

This extension helps pi use OpenAI Codex in a faster, lower-verbosity mode.

It is intended for openai-codex with gpt-5.5, the supported gpt-5.6 variants, gpt-6-astra, gpt-6-luna, gpt-6-sol, and gpt-6.1-sol, when you want quick execution and shorter responses.

Install

pi install npm:@ryan_nookpi/pi-extension-codex-fast-mode

Great for

• prioritizing speed over long explanations
• keeping Codex responses concise
• toggling a faster Codex setup per session

Usage

/codex-fast on
/codex-fast off
/codex-fast status

Notes

• Target models: openai-codex / gpt-5.5, gpt-5.6-luna, gpt-5.6-sol, gpt-5.6-terra, gpt-6-astra, gpt-6-luna, gpt-6-sol, and gpt-6.1-sol. gpt-5.4 is no longer targeted.
• Command descriptions and status messages abbreviate the list as gpt-5.5 ~ gpt-5.6, gpt-6, gpt-6.1-sol; matching still uses the exact model IDs above.
• This extension does not register models. All target models are in Pi SDK 1.0.1; availability still depends on your account and workspace.
• OpenAI documents Fast mode usage at 2.5x the Standard rate for included subscription limits and 2x for purchased credits and Enterprise pay-as-you-go usage. These are billing multipliers, not speed increases.
• GPT-5.5 retires from Codex on October 14, 2026; the OpenAI API is not affected.
• It always applies text.verbosity=low.
• When fast mode is enabled, it also injects service_tier=priority into the ChatGPT Codex backend request. Do not substitute the Codex CLI config value fast without verifying backend support.
• Live requests with Pi SDK 1.0.1 succeeded for all four targeted GPT-6 models with priority and low verbosity. The backend returned service_tier=default, so this confirms request acceptance, not actual priority scheduling or a speed increase.
• The setting is stored locally and persists across sessions.

References

• OpenAI documents GPT-6 Astra Fast mode availability and Codex credit consumption: https://developers.openai.com/codex/speed

• OpenAI documents Astra and its API Fast mode pricing: https://developers.openai.com/api/docs/models/gpt-6-astra

• Pi's model catalog and provider implementation: https://github.com/earendil-works/pi/tree/main/packages/ai/src/providers

• OpenAI recommends the Responses API for reasoning models such as gpt-5.6: https://developers.openai.com/api/docs/guides/text

• OpenAI documents text.verbosity=low for shorter GPT-5-family outputs: https://help.openai.com/en/articles/5072518-controlling-the-length-of-completions

• OpenAI documents service_tier=priority as the request-level opt-in for Priority processing on the Responses API: https://developers.openai.com/api/docs/guides/priority-processing