diff --git a/src/opengradient/client/llm.py b/src/opengradient/client/llm.py
index 20e32fb..20a570e 100644
--- a/src/opengradient/client/llm.py
+++ b/src/opengradient/client/llm.py
@@ -53,6 +53,11 @@ class _ChatParams:
     x402_settlement_mode: x402SettlementMode
 
 
+def _model_id(model: str) -> str:
+    """Extract model ID from a TEE_LLM value, handling plain strings without separator."""
+    return model.split("/")[1] if "/" in model else model
+
+
 class LLM:
     """
     LLM inference namespace.
@@ -277,7 +282,7 @@ async def completion(
         Raises:
             RuntimeError: If the inference fails.
         """
-        model_id = model.split("/")[1]
+        model_id = _model_id(model)
         payload: Dict = {
             "model": model_id,
             "prompt": prompt,
@@ -360,7 +365,7 @@ async def chat(
             RuntimeError: If the inference fails.
         """
         if response_format is not None and response_format.type == "json_object":
-            provider = model.split("/")[0]
+            provider = model.split("/")[0] if "/" in model else None
             if provider == "anthropic":
                 raise ValueError(
                     "Anthropic models do not support response_format type 'json_object'. "
@@ -368,7 +373,7 @@ async def chat(
                 )
 
         params = _ChatParams(
-            model=model.split("/")[1],
+            model=_model_id(model),
             max_tokens=max_tokens,
             temperature=temperature,
             stop_sequence=stop_sequence,