Merge pull request #38692 from BerriAI/litellm_model_info_parallel_fn_calling

fix(model_info): map supports_parallel_function_calling into ModelInfo

litellm/utils.py

@@ -5866,6 +5866,7 @@ def _get_model_info_helper(
                 supports_response_schema=_model_info.get("supports_response_schema", None),
                 supports_vision=_model_info.get("supports_vision", None),
                 supports_function_calling=_model_info.get("supports_function_calling", None),
+                supports_parallel_function_calling=_model_info.get("supports_parallel_function_calling", None),
                 supports_tool_choice=_model_info.get("supports_tool_choice", None),
                 supports_assistant_prefill=_model_info.get("supports_assistant_prefill", None),
                 supports_prompt_caching=_model_info.get("supports_prompt_caching", None),

tests/logging_callback_tests/gcs_pub_sub_body/spend_logs_payload.json

@@ -11,7 +11,7 @@
     "user": "",
     "team_id": "",
     "organization_id": "",
-    "metadata": "{\"applied_guardrails\": [], \"attempted_fallbacks\": null, \"original_model_group\": null, \"batch_models\": null, \"mcp_tool_call_metadata\": null, \"vector_store_request_metadata\": null, \"routing_decision\": null, \"internal_call_origin\": null, \"guardrail_information\": null, \"compression_savings\": null, \"usage_object\": {\"completion_tokens\": 20, \"prompt_tokens\": 10, \"total_tokens\": 30, \"completion_tokens_details\": null, \"prompt_tokens_details\": null}, \"model_map_information\": {\"model_map_key\": \"gpt-4o\", \"model_map_value\": {\"key\": \"gpt-4o\", \"max_tokens\": 16384, \"max_input_tokens\": 128000, \"max_output_tokens\": 16384, \"input_cost_per_token\": 2.5e-06, \"cache_creation_input_token_cost\": null, \"cache_read_input_token_cost\": 1.25e-06, \"input_cost_per_character\": null, \"input_cost_per_token_above_128k_tokens\": null, \"input_cost_per_token_above_200k_tokens\": null, \"input_cost_per_query\": null, \"input_cost_per_second\": null, \"input_cost_per_audio_token\": null, \"input_cost_per_token_batches\": 1.25e-06, \"output_cost_per_token_batches\": 5e-06, \"output_cost_per_token\": 1e-05, \"output_cost_per_audio_token\": null, \"output_cost_per_character\": null, \"output_cost_per_token_above_128k_tokens\": null, \"output_cost_per_character_above_128k_tokens\": null, \"output_cost_per_token_above_200k_tokens\": null, \"output_cost_per_second\": null, \"output_cost_per_image\": null, \"output_vector_size\": null, \"litellm_provider\": \"openai\", \"mode\": \"chat\", \"supports_system_messages\": true, \"supports_response_schema\": true, \"supports_vision\": true, \"supports_function_calling\": true, \"supports_tool_choice\": true, \"supports_assistant_prefill\": false, \"supports_prompt_caching\": true, \"supports_audio_input\": false, \"supports_audio_output\": false, \"supports_pdf_input\": false, \"supports_embedding_image_input\": false, \"supports_native_streaming\": null, \"supports_web_search\": true, \"supports_reasoning\": false, \"search_context_cost_per_query\": {\"search_context_size_low\": 0.03, \"search_context_size_medium\": 0.035, \"search_context_size_high\": 0.05}, \"tpm\": null, \"rpm\": null, \"supported_openai_params\": [\"frequency_penalty\", \"logit_bias\", \"logprobs\", \"top_logprobs\", \"max_tokens\", \"max_completion_tokens\", \"modalities\", \"prediction\", \"n\", \"presence_penalty\", \"seed\", \"stop\", \"stream\", \"stream_options\", \"temperature\", \"top_p\", \"tools\", \"tool_choice\", \"function_call\", \"functions\", \"max_retries\", \"extra_headers\", \"parallel_tool_calls\", \"audio\", \"response_format\", \"user\"]}}, \"additional_usage_values\": {\"completion_tokens_details\": null, \"prompt_tokens_details\": null}, \"user_api_key\": null, \"user_api_key_alias\": null, \"user_api_key_team_id\": null, \"user_api_key_project_id\": null, \"user_api_key_project_alias\": null, \"user_api_key_org_id\": null, \"user_api_key_user_id\": null, \"user_api_key_team_alias\": null, \"spend_logs_metadata\": null, \"requester_ip_address\": null, \"status\": null, \"proxy_server_request\": null, \"error_information\": null, \"attempted_retries\": null, \"max_retries\": null}",
+    "metadata": "{\"applied_guardrails\": [], \"attempted_fallbacks\": null, \"original_model_group\": null, \"batch_models\": null, \"mcp_tool_call_metadata\": null, \"vector_store_request_metadata\": null, \"routing_decision\": null, \"internal_call_origin\": null, \"guardrail_information\": null, \"compression_savings\": null, \"litellm_gateway_injected_cache\": null, \"usage_object\": {\"completion_tokens\": 20, \"prompt_tokens\": 10, \"total_tokens\": 30, \"completion_tokens_details\": null, \"prompt_tokens_details\": null}, \"model_map_information\": {\"model_map_key\": \"gpt-4o\", \"model_map_value\": {\"key\": \"gpt-4o\", \"max_tokens\": 16384, \"max_input_tokens\": 128000, \"max_output_tokens\": 16384, \"input_cost_per_token\": 2.5e-06, \"cache_creation_input_token_cost\": null, \"cache_read_input_token_cost\": 1.25e-06, \"input_cost_per_character\": null, \"input_cost_per_token_above_128k_tokens\": null, \"input_cost_per_token_above_200k_tokens\": null, \"input_cost_per_query\": null, \"input_cost_per_second\": null, \"input_cost_per_audio_token\": null, \"input_cost_per_token_batches\": 1.25e-06, \"output_cost_per_token_batches\": 5e-06, \"output_cost_per_token\": 1e-05, \"output_cost_per_audio_token\": null, \"output_cost_per_character\": null, \"output_cost_per_token_above_128k_tokens\": null, \"output_cost_per_character_above_128k_tokens\": null, \"output_cost_per_token_above_200k_tokens\": null, \"output_cost_per_second\": null, \"output_cost_per_image\": null, \"output_vector_size\": null, \"litellm_provider\": \"openai\", \"mode\": \"chat\", \"supports_system_messages\": true, \"supports_response_schema\": true, \"supports_vision\": true, \"supports_function_calling\": true, \"supports_tool_choice\": true, \"supports_assistant_prefill\": false, \"supports_prompt_caching\": true, \"supports_audio_input\": false, \"supports_audio_output\": false, \"supports_pdf_input\": false, \"supports_embedding_image_input\": false, \"supports_native_streaming\": null, \"supports_web_search\": true, \"supports_reasoning\": false, \"search_context_cost_per_query\": {\"search_context_size_low\": 0.03, \"search_context_size_medium\": 0.035, \"search_context_size_high\": 0.05}, \"tpm\": null, \"rpm\": null, \"supported_openai_params\": [\"frequency_penalty\", \"logit_bias\", \"logprobs\", \"top_logprobs\", \"max_tokens\", \"max_completion_tokens\", \"modalities\", \"prediction\", \"n\", \"presence_penalty\", \"seed\", \"stop\", \"stream\", \"stream_options\", \"temperature\", \"top_p\", \"tools\", \"tool_choice\", \"function_call\", \"functions\", \"max_retries\", \"extra_headers\", \"parallel_tool_calls\", \"audio\", \"response_format\", \"user\"]}}, \"additional_usage_values\": {\"completion_tokens_details\": null, \"prompt_tokens_details\": null}, \"user_api_key\": null, \"user_api_key_alias\": null, \"user_api_key_team_id\": null, \"user_api_key_project_id\": null, \"user_api_key_project_alias\": null, \"user_api_key_org_id\": null, \"user_api_key_user_id\": null, \"user_api_key_team_alias\": null, \"spend_logs_metadata\": null, \"requester_ip_address\": null, \"status\": null, \"proxy_server_request\": null, \"error_information\": null, \"attempted_retries\": null, \"max_retries\": null}",
     "cache_key": "Cache OFF",
     "spend": 0.00022500000000000002,
     "total_tokens": 30,

tests/test_litellm/proxy/proxy_server/test_routes_model_info.py

@@ -128,6 +128,20 @@ def test_v1_model_info_no_model_list_error(client, auth_as, null_router, path):
     assert "LLM Model List not loaded" in response.text
 
 
+
+def test_get_proxy_model_info_surfaces_supports_parallel_function_calling(local_model_cost_map):
+    """``GET /v1/model/info`` enriches each deployment through ``_get_proxy_model_info``; a registry
+    entry declaring parallel function calling must land in ``model_info`` instead of null."""
+    enriched = proxy_server._get_proxy_model_info(
+        model={
+            "model_name": "glm-5.3-flash",
+            "litellm_params": {"model": "together_ai/zai-org/GLM-5.3-Flash"},
+            "model_info": {"id": "glm-deployment", "db_model": False},
+        }
+    )
+    assert enriched["model_info"]["supports_parallel_function_calling"] is True
+
+
 def test_v1_model_info_star_wildcard_filter_keeps_provider_expansion(monkeypatch):
     from litellm.proxy._types import SpecialModelNames, UserAPIKeyAuth
     from litellm.proxy.auth import model_checks

tests/test_litellm/test_router.py

@@ -9051,6 +9051,25 @@ def test_model_group_info_reasoning_efforts_ignore_a_deployment_off_the_map():
     assert result.supported_reasoning_efforts == ("none", "minimal", "low", "medium", "high", "max")
 
 
+
+def test_model_group_info_surfaces_supports_parallel_function_calling(local_model_cost_map):
+    """``/model_group/info`` folds each deployment's registry flags into the group; a deployment whose
+    registry entry declares parallel function calling must flip the group to True instead of False."""
+    router = litellm.Router(
+        model_list=[
+            {
+                "model_name": "glm-group",
+                "litellm_params": {"model": "together_ai/zai-org/GLM-5.3-Flash", "api_key": "fake-key"},
+            }
+        ]
+    )
+
+    result = router._set_model_group_info(model_group="glm-group", user_facing_model_group_name="glm-group")
+
+    assert result is not None
+    assert result.supports_parallel_function_calling is True
+
+
 def test_model_group_info_reasoning_efforts_empty_on_a_mapped_non_reasoning_deployment():
     """A group mixing a reasoning model with one the map knows is not a reasoning model shares no
     level, so it advertises none and the picker offers nothing rather than a level routing would

tests/test_litellm/test_utils.py

@@ -120,6 +120,21 @@ def test_get_model_info_surfaces_supports_adaptive_thinking(local_model_cost_map
     assert generalized["supports_adaptive_thinking"] is True
 
 
+
+def test_get_model_info_surfaces_supports_parallel_function_calling(local_model_cost_map):
+    """A registry entry's supports_parallel_function_calling must read back through get_model_info
+    and litellm.supports_parallel_function_calling. Regression: the key was never copied into
+    ModelInfo, so provider-prefixed entries read None / False even when the map said True, and an
+    explicit False was indistinguishable from unset."""
+    declared_true = litellm.get_model_info(model="together_ai/zai-org/GLM-5.3-Flash")
+    assert declared_true["supports_parallel_function_calling"] is True
+    assert litellm.supports_parallel_function_calling(model="together_ai/zai-org/GLM-5.3-Flash") is True
+
+    declared_false = litellm.get_model_info(model="o3-mini")
+    assert declared_false["supports_parallel_function_calling"] is False
+    assert litellm.supports_parallel_function_calling(model="o3-mini") is False
+
+
 def test_get_model_info_surfaces_supported_endpoints(local_model_cost_map):
     """supported_endpoints ships in the cost map and is declared on ModelInfoBase,
     but the constructor never copied it, so get_model_info always returned None.