[ENHANCEMENT] GPT 5.4 FAST MODE CODEX

[Fernado03]

Problem (one or two sentences)

Add the Fast mode in gpt-5.4 and gpt-5.4-mini

Context (who is affected and when)

by default, to use this on codex, the user can trigger /fast

Desired behavior (conceptual, not technical)

faster agent responses

Constraints / preferences (optional)

No response

Request checklist

• I've searched existing Issues and Discussions for duplicates
• This describes a specific problem with clear context and impact

Roo Code Task Links (optional)

No response

Acceptance criteria (optional)

No response

Proposed approach (optional)

No response

Trade-offs / risks (optional)

No response

[roomote-v0]

I'd like to help with this issue, but I need some clarification to ensure I implement the right solution. Could you please provide more details on the following:

• What exactly is "Fast mode"? The codebase already includes gpt-5.4, gpt-5.4-mini, and gpt-5.4-nano models for both the OpenAI Native and OpenAI Codex providers. There is a "spark" variant (gpt-5.3-codex-spark) that is described as a "fast, text-only" model. Are you asking for a similar spark/fast variant for GPT-5.4?

• Is this a new model ID from OpenAI? For example, is there a new gpt-5.4-codex-spark or gpt-5.4-fast model that OpenAI has released that needs to be added? If so, could you share the model specifications (context window, max tokens, supported features)?

• Or is this about a configuration/reasoning effort setting? The existing GPT-5.4 models already support reasoning effort levels from "none" to "xhigh". Setting reasoning effort to "none" or "low" would give faster responses. Is the request about exposing a convenient "/fast" command that adjusts these settings?

• What does "/fast" refer to specifically? Is this a ChatGPT Codex CLI feature, and if so, what model ID does it map to?

Without this context, it is hard to determine what changes are needed since the GPT-5.4 model family is already well-supported in the codebase.

Please tag @roomote in your response with the answers, and I'll be happy to implement the fix once I have this information.

[Fernado03]

https://developers.openai.com/codex/speed

hopefully this will help

@roomote

[roomote-v0]

Thanks for the link, @Fernado03! I've seen your comment and will explore adding GPT-5.4 fast/spark model variants for the Codex provider, following the existing gpt-5.3-codex-spark pattern. I'll open a PR to share progress.

[roomote-v0]

Thanks for the link, @Fernado03! I've reviewed the Codex speed documentation and the existing model definitions. I'll add GPT-5.4 Codex and GPT-5.4 Codex Spark (fast mode) model entries to the OpenAI Codex provider, following the same pattern as the existing gpt-5.3-codex and gpt-5.3-codex-spark models. I'll open a PR to share progress shortly.

[roomote-v0]

Opened a draft PR with my attempt at addressing this: #12113

Added gpt-5.4-codex-spark and gpt-5.4-mini-codex-spark fast model variants to the Codex provider, following the existing gpt-5.3-codex-spark pattern. Feedback welcome!

[roomote-v0]

Opened a draft PR with my attempt at addressing this: #12114

Added gpt-5.4-codex, gpt-5.4-codex-spark, and gpt-5.4-mini-codex-spark model entries to the OpenAI Codex provider. The spark variants are the "fast mode" models with reduced context/output limits and text-only support, following the same pattern as gpt-5.3-codex-spark. Feedback welcome!