Litellm dev 02 07 2025 p3 (#8387)

* add back streaming for base o3 (#8361) * test(base_llm_unit_tests.py): add base test for o-series models - ensure streaming always works * fix(base_llm_unit_tests.py): fix test for o series models * refactor: move test --------- Co-authored-by: Matteo Boschini <[email protected]>
BerriAI · Feb 8, 2025 · c83498f · c83498f
1 parent b242c66
commit c83498f
Show file tree

Hide file tree

Showing 2 changed files with 24 additions and 1 deletion.
diff --git a/litellm/llms/openai/chat/o_series_transformation.py b/litellm/llms/openai/chat/o_series_transformation.py
@@ -54,7 +54,7 @@ def should_fake_stream(
 
         if model is None:
             return True
-        supported_stream_models = ["o1-mini", "o1-preview"]
+        supported_stream_models = ["o1-mini", "o1-preview", "o3-mini"]
         for supported_model in supported_stream_models:
             if supported_model in model:
                 return False

diff --git a/tests/llm_translation/test_openai_o1.py b/tests/llm_translation/test_openai_o1.py
@@ -204,3 +204,26 @@ def test_o3_reasoning_effort():
         reasoning_effort="high",
     )
     assert resp.choices[0].message.content is not None
+
+
+def test_streaming_response():
+    """Test that streaming response is returned correctly"""
+    from litellm import completion
+
+    response = completion(
+        model="o3-mini",
+        messages=[
+            {"role": "system", "content": "Be a good bot!"},
+            {"role": "user", "content": "Hello!"},
+        ],
+        stream=True,
+    )
+
+    assert response is not None
+
+    chunks = []
+    for chunk in response:
+        chunks.append(chunk)
+
+    resp = litellm.stream_chunk_builder(chunks=chunks)
+    print(resp)