sync(openai): remove gpt-image-1.5, document gpt-image-2 multi-res sizes

[Copilot]

Upstream Docs (f87f857) dropped gpt-image-1.5 and expanded gpt-image-2 to support 5 aspect ratios × 3 resolution tiers (1K/2K/4K) with differentiated billing. Local docs were out of sync on both fronts.

Changes

Remove gpt-image-1.5

• Dropped from intro, conclusion, and README model tag list in both openai_images_generations and openai_images_edits guides — model is no longer offered via the API.

Document gpt-image-2 high-resolution support

• Updated the "Native multi-aspect-ratio" bullet to reflect 5 ratios × 3 tiers (1K/2K/4K).
• Added "Supported size Values and Billing Tiers" table to both guides:

Aspect Ratio	1K (standard)	2K (×1.5)	4K (×1.5)
1:1	1024x1024	2048x2048	2880x2880
4:3	1536x1024	2048x1536	3264x2448
3:4	1024x1536	1536x2048	2448x3264
16:9	1792x1024	2048x1152	3840x2160
9:16	1024x1792	1152x2048	2160x3840

• Added note that 4K calls typically take 4–8 min and should use callback_url.
• Added "Supports high-resolution redraw" bullet to the edits guide (pass 1K source, get 2K/4K output in one call).

[acedatacloud-dev]

Auto-closed during PR backlog cleanup: superseded sync (+34/-6) — content covered by next dispatch against current Docs. Root cause was a broken auto-merge in sync-from-docs.yml (now fixed in PR #143). The fix lands a pre-cleanup step + reliable completion signal so future dispatches will not leak PRs. If this PR contained content not yet in main, the next docs-updated dispatch will re-create it via the working pipeline.