### Check for existing issues

- [x] I have searched the existing issues and checked that my issue is not a duplicate.

### What happened?

the first few characters from anthropic models are trunacted when calling litellm responses api. run the script below to reporudce, we expect I'm happy, but we get 'm happy

### Steps to Reproduce

import litellm

response = litellm.responses( model="anthropic/claude-sonnet-4-5", input=[{ "type": "message", "role": "user", "content": [{"type": "text", "text": "repeat after me: I'm happy to be here"}] }], max_tokens=100, stream=True )

for event in response: if hasattr(event, 'delta') and event.delta: print(event.delta)

### Relevant log output

```shell

```

### What part of LiteLLM is this about?

SDK (litellm Python package)

### What LiteLLM version are you on ?

1.81.5

### Twitter / LinkedIn details

_No response_