fix(core): sanitize Unicode surrogates in all LLM client requests (#9323)

Multiple OpenAI-compatible LLM clients (Azure, Deepseek, Groq, Together, XAI, ZAI)
and Anthropic-compatible clients (Anthropic, MiniMax, Google Vertex) were overriding
request_async/stream_async without calling sanitize_unicode_surrogates, causing
UnicodeEncodeError when message content contained lone UTF-16 surrogates.

Root cause: Child classes override parent methods but omit the sanitization step that
the base OpenAIClient includes. This allows corrupted Unicode (unpaired surrogates
from malformed emoji) to reach the httpx layer, which rejects it during UTF-8 encoding.

Fix: Import and call sanitize_unicode_surrogates in all overridden request methods.
Also removed duplicate sanitize_unicode_surrogates definition from openai_client.py
that shadowed the canonical implementation in letta.helpers.json_helpers.

🐾 Generated with [Letta Code](https://letta.com)

Co-Authored-By: Letta <noreply@letta.com>

Issue-ID: 10c0f2e4-f87b-11f0-b91c-da7ad0900000

letta/llm_api/azure_client.py

@@ -4,6 +4,7 @@ from typing import List, Optional, Tuple
 from openai import AsyncAzureOpenAI, AzureOpenAI
 from openai.types.chat.chat_completion import ChatCompletion
 
+from letta.helpers.json_helpers import sanitize_unicode_surrogates
 from letta.llm_api.openai_client import OpenAIClient
 from letta.otel.tracing import trace_method
 from letta.schemas.embedding_config import EmbeddingConfig
@@ -49,6 +50,8 @@ class AzureClient(OpenAIClient):
         """
         Performs underlying asynchronous request to OpenAI API and returns raw response dict.
         """
+        request_data = sanitize_unicode_surrogates(request_data)
+
         api_key, base_url, api_version = await self.get_byok_overrides_async(llm_config)
         if not api_key or not base_url or not api_version:
             api_key = model_settings.azure_api_key or os.environ.get("AZURE_API_KEY")