[Model Routing] claude-sonnet-5 swallows streaming assistant.reasoning_delta; flushes all thinkingContent at completion

Environment Details:

Working Model: claude-sonnet-4.5 (streams perfectly)

Failing Model: claude-sonnet-5

Integration: GitHub Copilot SDK (subscribe('assistant.reasoning_delta'))

Observed Behavior:
When routing requests to claude-sonnet-5, the SDK emits zero assistant.reasoning_delta events during the generation lifecycle. Instead, it accumulates them in the background, resulting in a ~25,000ms Time-to-First-Token (TTFT) and blocking client-side UI rendering. A complete, non-empty thinkingContent block is delivered atomically only when the final text stream buffer flushes or the file completes.

Expected Behavior:
The SDK should map Anthropic's streaming reasoning frames into real-time assistant.reasoning_delta events immediately upon receiving them from the upstream model router, preserving low operational TTFT just as it successfully does for claude-sonnet-4.5.

从 subscribe('assistant.reasoning_delta') 背后的实现开始，跟踪 claude-sonnet-5 的模型路由路径。将其处理方式与 claude-sonnet-4.5 进行比较，并验证上游推理帧是否会立即发出，而不是进行缓冲。完成标准是 claude-sonnet-5 能够实时发出 reasoning_delta 事件并保持较低的 TTFT，而无需等待最终完成。

由索引模型根据 Issue 内容生成。