feat: Add GPT-5 Pro with background mode auto-resume and polling

[hannesrudolph]

Summary

Adds GPT-5 Pro model with OpenAI Responses API background mode support. Background requests can take several minutes, so this implements resilient streaming with automatic recovery.

Why

GPT-5 Pro is a slow, reasoning-focused model that can take several minutes to respond. The standard streaming approach times out or appears stuck. OpenAI's Responses API background mode is designed for these long-running requests.

What Changed

Model Addition

• Added gpt-5-pro-2025-10-06 with backgroundMode: true flag in model metadata

Background Mode Implementation

• Auto-enables background: true, stream: true, store: true for flagged models
• Emits status events: queued → in_progress → completed/failed
• Shows status labels in UI spinner ("background mode (queued)…", etc.)

Resilient Streaming

• Auto-resume: If stream drops, resumes from last sequence number using GET /v1/responses/{id}?starting_after={seq}
• Poll fallback: If resume fails after 3 retries, polls every 2s until completion (up to 20 minutes)
• Synthesizes final output and usage data when polling completes

Files Changed

• packages/types/src/providers/openai.ts - Model metadata
• src/api/providers/openai-native.ts - Background mode logic, auto-resume, polling
• src/core/task/Task.ts - Status event handling
• webview-ui/src/utils/backgroundStatus.ts - Status label mapping
• webview-ui/src/components/chat/* - UI status display

Testing

• Background mode status emission and lifecycle
• Auto-resume on stream drop with exponential backoff
• Poll fallback when resume exhausts retries
• Usage tracking parity with non-background requests
• UI status label mapping

---

Important

Adds GPT-5 Pro model with background mode, implementing resilient streaming, auto-resume, and UI status updates.

• Model Addition:
  • Added gpt-5-pro-2025-10-06 with backgroundMode: true in openai.ts.
• Background Mode Implementation:
  • Auto-enables background: true, stream: true, store: true for flagged models in openai-native.ts.
  • Emits status events: queued → in_progress → completed/failed.
  • Shows status labels in UI spinner in ChatRow.tsx and ChatView.tsx.
• Resilient Streaming:
  • Auto-resume: Resumes from last sequence number using GET /v1/responses/{id}?starting_after={seq} in openai-native.ts.
  • Poll fallback: Polls every 2s until completion if resume fails, up to 20 minutes.
• Files Changed:
  • openai.ts, openai-native.ts, Task.ts - Background mode logic, auto-resume, polling.
  • ChatRow.tsx, ChatView.tsx - UI status display.
  • backgroundStatus.ts, backgroundStatus.spec.ts - Status label mapping.
• Testing:
  • Background mode status emission and lifecycle.
  • Auto-resume on stream drop with exponential backoff.
  • Poll fallback when resume exhausts retries.
  • Usage tracking parity with non-background requests.
  • UI status label mapping.

^{This description was created by for 3a0add7. You can customize this summary. It will automatically update as commits are pushed.}

[hannesrudolph]

Implemented minor fixes and a small performance improvement based on the review. See inline notes.

[roomote]

Status: Posted targeted inline review comments. Focused on metadata consistency and text style. Background-mode implementation (auto-resume/poll) appears solid with explicit logging and error classification.

• Resolve reasoning effort config mismatch for GPT-5 Pro model metadata (supportsReasoningEffort is false while a default reasoningEffort of "high" is set). Decide whether to remove the default or enable configurability.
• Replace em dash in GPT‑5 Pro description with the standard hyphen form to match existing UI copy style.
• Verified background-mode resilience and logging (resume backoff, polling classification) in provider path.

_{Mention @roomote in a comment to trigger your PR Fixer agent and make changes to this pull request.}

[roomote]

Reasoning effort config is contradictory here: supportsReasoningEffort is false but reasoningEffort is set to "high". With this combination the UI hides the selector while the backend still injects a reasoning parameter (see src/shared/api.ts and src/api/providers/openai-native.ts). This can be confusing for users and maintainers. Consider either enabling supportsReasoningEffort to reflect configurability or removing the default reasoningEffort and documenting that the model runs with provider defaults.

_{Fix it with Roo Code or mention @roomote and request a fix.}

[roomote]

Style nit: The description uses an em dash (—). Project text typically avoids em dashes for consistency with UI strings. Consider replacing with a spaced hyphen form to match style elsewhere, e.g.:

"GPT-5 Pro: A slow, reasoning-focused model for complex problems. Uses background mode with resilient streaming - requests may take some time and will automatically reconnect if they time out."

_{Fix it with Roo Code or mention @roomote and request a fix.}