## Description

When using LiteLLM's OpenAI-compatible proxy (`/v1/chat/completions`) with a Claude model and the **Vercel AI SDK 6.x** (`ai@6.x`), multi-step tool calls produce streaming errors of the form:

``` text part chatcmpl-<id> not found ```

This causes the AI SDK to silently **drop text deltas** for the affected steps — producing empty or truncated assistant responses.

## Environment

- LiteLLM version: 1.x (confirmed on hosted proxy at `llm.behaviorlabs.ai/v1`) - Model: `claude-sonnet-4-6` (via LiteLLM Claude alias) - Client: `@ai-sdk/openai@3.x` pointing at LiteLLM proxy - AI SDK: `ai@6.0.105` - Framework: Next.js App Router, `streamText` + `stopWhen: stepCountIs(20)` for multi-step tool use

## Steps to Reproduce

1. Configure `@ai-sdk/openai` to point at LiteLLM proxy: ```ts const litellm = createOpenAI({ baseURL: 'https://<litellm>/v1', apiKey: '...' }) ``` 2. Call `streamText` with tools and `stopWhen: stepCountIs(20)` using a Claude model: ```ts const result = streamText({ model: litellm('claude-sonnet-4-6'), tools: { searchEvidence: ..., webSearch: ... }, stopWhen: stepCountIs(20), }) ``` 3. Ask a question that triggers tool u…