[ENHANCEMENT] Add 1.5x Fast mode for the OpenAI ChatGPT subscription provider

[WebMad]

Problem (one or two sentences)

Users of the OpenAI ChatGPT Plus/Pro subscription provider cannot enable Codex priority processing from Zoo Code. As a result, they cannot opt into the approximately 1.5x faster response mode available for subscription-backed Codex requests.

Context (who is affected and when)

This affects users authenticated through the OpenAI ChatGPT Plus/Pro subscription provider, especially when running latency-sensitive coding tasks where faster model responses are more important than conserving subscription quota.

Desired behavior (conceptual, not technical)

Allow users to choose between Standard and Fast processing in the OpenAI ChatGPT subscription provider settings. Fast mode should clearly indicate that it provides approximately 1.5x speed and consumes more subscription quota, while Standard remains the default.

Constraints / preferences (optional)

• Keep Standard as the default so existing users' behavior and quota consumption do not change.
• Clearly communicate the speed/quota trade-off before Fast mode is selected.
• Persist the selection as part of the provider configuration.
• Apply the choice consistently to streaming and non-streaming requests.

Request checklist

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

Zoo Code Task Links (optional)

N/A

---

Acceptance criteria (optional)

Given a user authenticated with the OpenAI ChatGPT subscription provider
When they open the provider settings
Then they can choose between Standard and Fast processing
And Fast is described as approximately 1.5x speed with increased subscription usage
And Standard is selected by default when no preference has been saved

Given Fast processing is selected
When Zoo Code sends either a streaming or non-streaming Codex request
Then priority processing is requested

Given Standard processing is selected
When Zoo Code sends a Codex request
Then priority processing is not requested

Proposed approach (optional)

Add a Speed selector to the OpenAI ChatGPT subscription provider with Standard and Fast options. Map Fast to the Codex priority service tier and include it in both streaming and non-streaming Responses API requests.

Trade-offs / risks (optional)

Fast mode consumes subscription quota more quickly. The upstream service may also change availability, behavior, or quota accounting for priority processing, so the UI should avoid promising an exact latency improvement.

[WebMad]

Upstream implementation references

There does not appear to be a standalone OpenAI documentation article for this Codex subscription option. The mapping is confirmed by OpenAI's official Codex repository:

• Codex model catalog at an immutable commit
• Current model catalog on main
• Official Responses WebSocket request serialization

The model catalog declares the option as:

{
  "id": "priority",
  "name": "Fast",
  "description": "1.5x speed, increased usage"
}

The official Codex request serialization uses service_tier: Some("priority".to_string()). Therefore the wire-level mapping for Zoo Code is:

• Fast → send service_tier: "priority"
• Standard → omit service_tier

No separate quota multiplier is sent by the client. The increased subscription usage is accounted for and enforced by OpenAI's server.

[WebMad]

claiming