fix: validate model string format before splitting in LLM methods

src/opengradient/client/llm.py

@@ -251,8 +251,12 @@ async def completion(
 
         Raises:
             RuntimeError: If the inference fails.
+            ValueError: If the model string is not in "provider/model" format.
         """
-        model_id = model.split("/")[1]
+        parts = model.split("/")
+        if len(parts) < 2 or not parts[1]:
+            raise ValueError(f"Invalid model format {model!r}: expected 'provider/model' (e.g. 'openai/gpt-5')")
+        model_id = parts[1]
         payload: Dict = {
             "model": model_id,
             "prompt": prompt,
@@ -325,9 +329,13 @@ async def chat(
 
         Raises:
             RuntimeError: If the inference fails.
+            ValueError: If the model string is not in "provider/model" format.
         """
+        parts = model.split("/")
+        if len(parts) < 2 or not parts[1]:
+            raise ValueError(f"Invalid model format {model!r}: expected 'provider/model' (e.g. 'openai/gpt-5')")
         params = _ChatParams(
-            model=model.split("/")[1],
+            model=parts[1],
             max_tokens=max_tokens,
             temperature=temperature,
             stop_sequence=stop_sequence,

tests/client_test.py

@@ -190,3 +190,34 @@ def test_settlement_modes_values(self):
         assert x402SettlementMode.PRIVATE == "private"
         assert x402SettlementMode.BATCH_HASHED == "batch"
         assert x402SettlementMode.INDIVIDUAL_FULL == "individual"
+
+
+# --- Model Format Validation Tests ---
+
+
+class TestModelFormatValidation:
+    """Tests for model string validation in LLM.completion() and LLM.chat()."""
+
+    def test_completion_rejects_model_without_slash(self, mock_tee_registry):
+        llm = LLM(private_key=FAKE_PRIVATE_KEY)
+        with pytest.raises(ValueError, match="Invalid model format"):
+            import asyncio
+            asyncio.get_event_loop().run_until_complete(
+                llm.completion(model="no-slash-model", prompt="hi")
+            )
+
+    def test_completion_rejects_model_with_trailing_slash(self, mock_tee_registry):
+        llm = LLM(private_key=FAKE_PRIVATE_KEY)
+        with pytest.raises(ValueError, match="Invalid model format"):
+            import asyncio
+            asyncio.get_event_loop().run_until_complete(
+                llm.completion(model="openai/", prompt="hi")
+            )
+
+    def test_chat_rejects_model_without_slash(self, mock_tee_registry):
+        llm = LLM(private_key=FAKE_PRIVATE_KEY)
+        with pytest.raises(ValueError, match="Invalid model format"):
+            import asyncio
+            asyncio.get_event_loop().run_until_complete(
+                llm.chat(model="no-slash-model", messages=[{"role": "user", "content": "hi"}])
+            )