From 4c022a3089cbfbc377cc700db7ade6309f3b1e1f Mon Sep 17 00:00:00 2001 From: mateo-berri <277851410+mateo-berri@users.noreply.github.com> Date: Sat, 12 Sep 2026 23:34:12 -0700 Subject: [PATCH 1/4] feat(pricing): add azure gpt-chat-latest global and data zone rates --- ...odel_prices_and_context_window_backup.json | 111 ++++++++++++++++++ model_prices_and_context_window.json | 111 ++++++++++++++++++ .../llm_cost_calc/test_llm_cost_calc_utils.py | 31 +++++ .../test_litellm/test_model_prices_schema.py | 8 ++ 4 files changed, 261 insertions(+) diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json index 2220d0e1fe5..86605d77aab 100644 --- a/litellm/model_prices_and_context_window_backup.json +++ b/litellm/model_prices_and_context_window_backup.json @@ -7409,6 +7409,80 @@ "supports_web_search": true, "supports_xhigh_reasoning_effort": true }, + "azure/gpt-chat-latest": { + "cache_read_input_token_cost": 5e-07, + "deprecation_date": "2026-12-02", + "input_cost_per_token": 5e-06, + "litellm_provider": "azure", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 3e-05, + "reasoning_effort_levels": [ + "medium" + ], + "source": "https://azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true + }, + "azure/chat-latest": { + "cache_read_input_token_cost": 5e-07, + "deprecation_date": "2026-12-02", + "input_cost_per_token": 5e-06, + "litellm_provider": "azure", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 3e-05, + "reasoning_effort_levels": [ + "medium" + ], + "source": "https://azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true + }, "azure/us/gpt-5.6": { "cache_creation_input_token_cost": 6.875e-06, "cache_creation_input_token_cost_above_272k_tokens": 1.375e-05, @@ -7675,6 +7749,43 @@ "supports_web_search": true, "supports_xhigh_reasoning_effort": true }, + "azure/us/gpt-chat-latest": { + "cache_read_input_token_cost": 5.5e-07, + "deprecation_date": "2026-12-02", + "input_cost_per_token": 5.5e-06, + "litellm_provider": "azure", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 3.3e-05, + "reasoning_effort_levels": [ + "medium" + ], + "source": "https://azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true + }, "azure/eu/gpt-5.6": { "cache_creation_input_token_cost": 6.875e-06, "cache_creation_input_token_cost_above_272k_tokens": 1.375e-05, diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index 2220d0e1fe5..86605d77aab 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -7409,6 +7409,80 @@ "supports_web_search": true, "supports_xhigh_reasoning_effort": true }, + "azure/gpt-chat-latest": { + "cache_read_input_token_cost": 5e-07, + "deprecation_date": "2026-12-02", + "input_cost_per_token": 5e-06, + "litellm_provider": "azure", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 3e-05, + "reasoning_effort_levels": [ + "medium" + ], + "source": "https://azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true + }, + "azure/chat-latest": { + "cache_read_input_token_cost": 5e-07, + "deprecation_date": "2026-12-02", + "input_cost_per_token": 5e-06, + "litellm_provider": "azure", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 3e-05, + "reasoning_effort_levels": [ + "medium" + ], + "source": "https://azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true + }, "azure/us/gpt-5.6": { "cache_creation_input_token_cost": 6.875e-06, "cache_creation_input_token_cost_above_272k_tokens": 1.375e-05, @@ -7675,6 +7749,43 @@ "supports_web_search": true, "supports_xhigh_reasoning_effort": true }, + "azure/us/gpt-chat-latest": { + "cache_read_input_token_cost": 5.5e-07, + "deprecation_date": "2026-12-02", + "input_cost_per_token": 5.5e-06, + "litellm_provider": "azure", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 3.3e-05, + "reasoning_effort_levels": [ + "medium" + ], + "source": "https://azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true + }, "azure/eu/gpt-5.6": { "cache_creation_input_token_cost": 6.875e-06, "cache_creation_input_token_cost_above_272k_tokens": 1.375e-05, diff --git a/tests/test_litellm/litellm_core_utils/llm_cost_calc/test_llm_cost_calc_utils.py b/tests/test_litellm/litellm_core_utils/llm_cost_calc/test_llm_cost_calc_utils.py index cbe6fe198c9..3c6a13a7f01 100644 --- a/tests/test_litellm/litellm_core_utils/llm_cost_calc/test_llm_cost_calc_utils.py +++ b/tests/test_litellm/litellm_core_utils/llm_cost_calc/test_llm_cost_calc_utils.py @@ -2101,6 +2101,37 @@ def test_generic_cost_per_token_azure_gpt_6_astra_foundry_price_sheet( assert completion_cost == pytest.approx(zone_multiplier * output_multiplier * completion_tokens * 5e-5) +@pytest.mark.parametrize( + "model,input_rate,cache_read_rate,output_rate", + [ + ("azure/gpt-chat-latest", 5e-6, 5e-7, 3e-5), + ("azure/chat-latest", 5e-6, 5e-7, 3e-5), + ("azure/us/gpt-chat-latest", 5.5e-6, 5.5e-7, 3.3e-5), + ], +) +def test_generic_cost_per_token_azure_gpt_chat_latest_price_sheet( + _local_model_cost_map, model, input_rate, cache_read_rate, output_rate +): + """The Azure OpenAI price sheet lists GPT-Chat Latest at $5 input, $0.50 cached input and $30 output per 1M + tokens on Global, and $5.50, $0.55 and $33 on Data Zone. Foundry names the product gpt-chat-latest and the + OpenAI API names the same model chat-latest, so both spellings bill the Global sheet. + """ + prompt_tokens = 100000 + cached_tokens = 40000 + completion_tokens = 1000 + usage = Usage( + prompt_tokens=prompt_tokens, + completion_tokens=completion_tokens, + total_tokens=prompt_tokens + completion_tokens, + prompt_tokens_details=PromptTokensDetailsWrapper(cached_tokens=cached_tokens), + ) + + prompt_cost, completion_cost = generic_cost_per_token(model=model, usage=usage, custom_llm_provider="azure") + + assert prompt_cost == pytest.approx((prompt_tokens - cached_tokens) * input_rate + cached_tokens * cache_read_rate) + assert completion_cost == pytest.approx(completion_tokens * output_rate) + + def test_generic_cost_per_token_azure_ai_gpt_6_astra_flex_bills_the_standard_rate(_local_model_cost_map): usage = Usage(prompt_tokens=1000, completion_tokens=100, total_tokens=1100) diff --git a/tests/test_litellm/test_model_prices_schema.py b/tests/test_litellm/test_model_prices_schema.py index 0b9dbd23097..e562797fbe8 100644 --- a/tests/test_litellm/test_model_prices_schema.py +++ b/tests/test_litellm/test_model_prices_schema.py @@ -221,6 +221,14 @@ def test_chat_latest_declares_the_one_effort_openai_accepts(prices: dict): assert resolve_supported_reasoning_efforts(prices["chat-latest"], deployment_is_mapped=True) == ("medium",) +@pytest.mark.parametrize("key", ["azure/gpt-chat-latest", "azure/chat-latest", "azure/us/gpt-chat-latest"]) +def test_azure_gpt_chat_latest_declares_the_one_effort_azure_accepts(prices: dict, key: str): + """Azure answers every reasoning_effort on a gpt-chat-latest deployment except medium with + "Unsupported value ... Supported values are: 'medium'", the same fixed level OpenAI's chat-latest + carries, so the Foundry product name and the OpenAI API name both declare that one level.""" + assert resolve_supported_reasoning_efforts(prices[key], deployment_is_mapped=True) == ("medium",) + + BEDROCK_OPENAI_GPT_MARKERS: Final = ("openai.gpt-5.4", "openai.gpt-5.5", "openai.gpt-5.6", "openai.gpt-6-astra") BEDROCK_PROVIDERS: Final = frozenset(("bedrock", "bedrock_converse", "bedrock_mantle")) BEDROCK_ROW_PREFIXES: Final = ("bedrock_mantle/", "us.", "global.") From a70f9a17ccb2e4dae4bb7a90236d514a45bc9e5b Mon Sep 17 00:00:00 2001 From: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> Date: Mon, 14 Sep 2026 08:54:34 +0000 Subject: [PATCH 2/4] perf(logging): skip correlation contextvar stamping when request_correlation_in_logs is off Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> --- litellm/litellm_core_utils/litellm_logging.py | 8 +- .../test_litellm_logging.py | 90 +++++++++++++++++-- .../test_streaming_handler.py | 24 +++-- 3 files changed, 104 insertions(+), 18 deletions(-) diff --git a/litellm/litellm_core_utils/litellm_logging.py b/litellm/litellm_core_utils/litellm_logging.py index 5fe94320491..9ba9fd082f3 100644 --- a/litellm/litellm_core_utils/litellm_logging.py +++ b/litellm/litellm_core_utils/litellm_logging.py @@ -556,7 +556,7 @@ class Logging(LiteLLMLoggingBaseClass): # ids leaking into a different, later request on the same thread. Sync # support is deferred to a follow-up PR with its own safe-restore # mechanism; async calls (the proxy's only call path) are unaffected. - if supports_correlation_logging: + if supports_correlation_logging and litellm.request_correlation_in_logs: set_trace_id(self.litellm_trace_id) set_session_id(self.litellm_session_id) # set_trace_id()/set_session_id() sanitize (strip control chars, bound @@ -2442,7 +2442,7 @@ class Logging(LiteLLMLoggingBaseClass): call) would leave the outer request's subsequent log lines stamped with the nested call's trace_id/session_id instead of its own. - Uses a plain set() of the captured pre-call value rather than + Uses a plain contextvar set() of the captured pre-call value rather than contextvars.Token-based reset(), since this can end up called from a different asyncio Task/context than __init__ ran in (e.g. the request task's own wrapper() finally block, plus async_success_handler @@ -2453,8 +2453,8 @@ class Logging(LiteLLMLoggingBaseClass): that Task's view of the contextvars, so calling it multiple times (once per Task involved in this attempt) is required, not just safe. """ - set_trace_id(self._pre_call_trace_id) - set_session_id(self._pre_call_session_id) + trace_id_var.set(self._pre_call_trace_id) + session_id_var.set(self._pre_call_session_id) def _restore_correlation_context_if_unclaimed(self) -> None: """Guarded variant for __del__-triggered cleanup only. diff --git a/tests/test_litellm/litellm_core_utils/test_litellm_logging.py b/tests/test_litellm/litellm_core_utils/test_litellm_logging.py index 451b740dd71..70f9bae283b 100644 --- a/tests/test_litellm/litellm_core_utils/test_litellm_logging.py +++ b/tests/test_litellm/litellm_core_utils/test_litellm_logging.py @@ -5221,10 +5221,11 @@ def test_handle_anthropic_messages_parsed_response_logging_preserves_fast_mode_s assert getattr(result.usage, "speed", None) == "fast" -def test_logging_init_sets_trace_id(): +def test_logging_init_sets_trace_id(monkeypatch): """Logging.__init__() must call set_trace_id with self.litellm_trace_id.""" from litellm.litellm_core_utils.litellm_logging import Logging + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("") log_obj = Logging( @@ -5240,7 +5241,7 @@ def test_logging_init_sets_trace_id(): assert trace_id_var.get() == log_obj.litellm_trace_id -def test_logging_init_skips_stamping_when_correlation_logging_unsupported(): +def test_logging_init_skips_stamping_when_correlation_logging_unsupported(monkeypatch): """supports_correlation_logging=False (what wrapper(), the sync entry point, always passes) must leave trace_id_var/session_id_var completely untouched, even though self.litellm_trace_id/litellm_session_id (the @@ -5248,6 +5249,7 @@ def test_logging_init_skips_stamping_when_correlation_logging_unsupported(): usual - only the ambient contextvar stamping is gated.""" from litellm.litellm_core_utils.litellm_logging import Logging + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("") session_id_var.set("") @@ -5271,10 +5273,48 @@ def test_logging_init_skips_stamping_when_correlation_logging_unsupported(): assert log_obj.litellm_session_id == "should-not-be-stamped" -def test_logging_init_sets_session_id_when_provided(): +def test_logging_init_skips_stamping_when_request_correlation_in_logs_disabled(monkeypatch): + from litellm.litellm_core_utils.litellm_logging import Logging + + monkeypatch.setattr(litellm, "request_correlation_in_logs", False) + trace_id_var.set("outer") + session_id_var.set("outer-sid") + try: + with ( + patch( # test-quality-ok: regression test verifies disabled stamping skips both setters + "litellm.litellm_core_utils.litellm_logging.set_trace_id" + ) as mock_set_trace_id, + patch( # test-quality-ok: regression test verifies disabled stamping skips both setters + "litellm.litellm_core_utils.litellm_logging.set_session_id" + ) as mock_set_session_id, + ): + log_obj = Logging( + model="gpt-3.5-turbo", + messages=[{"role": "user", "content": "hi"}], + stream=False, + call_type="completion", + start_time=None, + litellm_call_id="call-disabled", + function_id="fn-disabled", + kwargs={"litellm_session_id": "disabled-session"}, + supports_correlation_logging=True, + ) + + assert trace_id_var.get() == "outer" + assert session_id_var.get() == "outer-sid" + assert log_obj._own_trace_id == "outer" + mock_set_trace_id.assert_not_called() + mock_set_session_id.assert_not_called() + finally: + trace_id_var.set("") + session_id_var.set("") + + +def test_logging_init_sets_session_id_when_provided(monkeypatch): """Logging.__init__() must call set_session_id when litellm_session_id is in kwargs.""" from litellm.litellm_core_utils.litellm_logging import Logging + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) session_id_var.set("") Logging( @@ -5290,11 +5330,12 @@ def test_logging_init_sets_session_id_when_provided(): assert session_id_var.get() == "my-session-99" -def test_logging_init_resets_session_id_to_empty_when_absent(): +def test_logging_init_resets_session_id_to_empty_when_absent(monkeypatch): """When no session_id is in kwargs, Logging.__init__() must reset session_id_var to "" so a prior request's session_id does not leak into subsequent log records.""" from litellm.litellm_core_utils.litellm_logging import Logging + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) session_id_var.set("preexisting-sid") Logging( @@ -5310,7 +5351,7 @@ def test_logging_init_resets_session_id_to_empty_when_absent(): assert session_id_var.get() == "" -def test_restore_correlation_context_resets_to_pre_call_value(): +def test_restore_correlation_context_resets_to_pre_call_value(monkeypatch): """_restore_correlation_context() must put trace_id_var/session_id_var back to whatever they were immediately before this Logging instance was constructed. This is the mechanism that prevents a nested call (e.g. a guardrail's own @@ -5318,6 +5359,7 @@ def test_restore_correlation_context_resets_to_pre_call_value(): session_id into the outer call's subsequent log lines.""" from litellm.litellm_core_utils.litellm_logging import Logging + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("outer-trace") session_id_var.set("outer-session") try: @@ -5343,7 +5385,7 @@ def test_restore_correlation_context_resets_to_pre_call_value(): session_id_var.set("") -def test_restore_correlation_context_safe_to_call_repeatedly(): +def test_restore_correlation_context_safe_to_call_repeatedly(monkeypatch): """Calling _restore_correlation_context() more than once must not raise. It's deliberately NOT guarded against repeat calls: wrapper()'s finally @@ -5353,6 +5395,7 @@ def test_restore_correlation_context_safe_to_call_repeatedly(): the contextvars, so repeat calls are expected, not just tolerated.""" from litellm.litellm_core_utils.litellm_logging import Logging + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) log_obj = Logging( model="gpt-3.5-turbo", messages=[{"role": "user", "content": "hi"}], @@ -5367,8 +5410,40 @@ def test_restore_correlation_context_safe_to_call_repeatedly(): log_obj._restore_correlation_context() # must not raise +def test_restore_correlation_context_does_not_resanitize(monkeypatch): + from litellm.litellm_core_utils.litellm_logging import Logging + from litellm._logging import _sanitize_correlation_id + + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) + trace_id_var.set("outer-trace") + session_id_var.set("outer-session") + try: + log_obj = Logging( + model="gpt-3.5-turbo", + messages=[{"role": "user", "content": "hi"}], + stream=False, + call_type="completion", + start_time=None, + litellm_call_id="call-no-resanitize", + function_id="fn-no-resanitize", + kwargs={"litellm_session_id": "inner-session"}, + ) + + with patch( # test-quality-ok: regression test verifies restore avoids sanitization + "litellm._logging._sanitize_correlation_id", wraps=_sanitize_correlation_id + ) as mock_sanitize: + log_obj._restore_correlation_context() + + mock_sanitize.assert_not_called() + assert trace_id_var.get() == "outer-trace" + assert session_id_var.get() == "outer-session" + finally: + trace_id_var.set("") + session_id_var.set("") + + @pytest.mark.asyncio -async def test_restore_correlation_context_works_across_asyncio_task_boundary(): +async def test_restore_correlation_context_works_across_asyncio_task_boundary(monkeypatch): """_restore_correlation_context() must succeed even when it's called from a different asyncio Task than the one Logging.__init__() ran in - exactly what happens on litellm's real async success path, where async_success_handler is @@ -5385,6 +5460,7 @@ async def test_restore_correlation_context_works_across_asyncio_task_boundary(): """ from litellm.litellm_core_utils.litellm_logging import Logging + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("outer-trace-cross-task") session_id_var.set("outer-session-cross-task") try: diff --git a/tests/test_litellm/litellm_core_utils/test_streaming_handler.py b/tests/test_litellm/litellm_core_utils/test_streaming_handler.py index 37e2031fdf4..784468c839b 100644 --- a/tests/test_litellm/litellm_core_utils/test_streaming_handler.py +++ b/tests/test_litellm/litellm_core_utils/test_streaming_handler.py @@ -4100,7 +4100,7 @@ async def test_async_streaming_completion_does_not_reset_context_before_iteratio session_id_var.set("") -def test_stream_wrapper_del_restores_correlation_context(): +def test_stream_wrapper_del_restores_correlation_context(monkeypatch): """CustomStreamWrapper.__del__ is the best-effort fallback for an abandoned stream (caller never exhausts it, so the normal terminal-handler restore never fires). Testing this via real garbage collection is unreliable in @@ -4112,6 +4112,7 @@ def test_stream_wrapper_del_restores_correlation_context(): doesn't run actual finalization, and this exercises exactly the logic that real garbage collection would eventually trigger. """ + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("outer-trace-abandoned") session_id_var.set("outer-session-abandoned") try: @@ -4159,12 +4160,13 @@ def test_stream_wrapper_del_never_raises_with_broken_logging_obj(): wrapper.__del__() # must not raise -def test_stream_wrapper_del_does_not_clobber_a_newer_active_call(): +def test_stream_wrapper_del_does_not_clobber_a_newer_active_call(monkeypatch): """A delayed finalizer must never stomp a different, still-active call's context. If an abandoned stream's __del__ fires late - after a new call has already started in the same Task/thread and claimed the contextvars - unconditionally restoring the abandoned stream's own pre-call snapshot would corrupt the active call's subsequent log lines with stale ids.""" + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("outer-trace-before-abandoned-call") session_id_var.set("outer-session-before-abandoned-call") try: @@ -4210,13 +4212,14 @@ def test_stream_wrapper_del_does_not_clobber_a_newer_active_call(): session_id_var.set("") -def test_stream_wrapper_del_restores_when_own_session_id_needed_sanitizing(): +def test_stream_wrapper_del_restores_when_own_session_id_needed_sanitizing(monkeypatch): """The __del__ guard must compare against the *sanitized* id actually stored in the contextvar, not the raw litellm_session_id/litellm_trace_id - set_session_id()/set_trace_id() strip control characters before storing, so a caller-supplied id containing e.g. a newline would never equal the raw attribute, and the guard would wrongly conclude some other call has claimed the context and skip cleanup forever.""" + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("outer-trace-needs-sanitizing") session_id_var.set("outer-session-needs-sanitizing") try: @@ -4250,7 +4253,7 @@ def test_stream_wrapper_del_restores_when_own_session_id_needed_sanitizing(): session_id_var.set("") -def test_stream_wrapper_next_keeps_context_active_through_synthesized_finish_reason_chunk(): +def test_stream_wrapper_next_keeps_context_active_through_synthesized_finish_reason_chunk(monkeypatch): """When the underlying stream ends without ever emitting an explicit finish_reason chunk, __next__ synthesizes one via finish_reason_handler() and returns it. That chunk is still this call's own data - the caller's @@ -4261,6 +4264,7 @@ def test_stream_wrapper_next_keeps_context_active_through_synthesized_finish_rea correct, deterministic restore on the very next __next__() call, since completion_stream is already exhausted and immediately re-raises StopIteration.""" + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("outer-trace-finish-reason") session_id_var.set("outer-session-finish-reason") try: @@ -4300,12 +4304,13 @@ def test_stream_wrapper_next_keeps_context_active_through_synthesized_finish_rea session_id_var.set("") -def test_stream_wrapper_del_cleans_up_after_synthesized_finish_reason_chunk(): +def test_stream_wrapper_del_cleans_up_after_synthesized_finish_reason_chunk(monkeypatch): """A caller that breaks immediately after seeing finish_reason (the early-break pattern) never triggers the next()-driven restore above - it relies on the best-effort __del__ guard instead, same as any other abandoned stream. The guard must still recognize this call's own (unrestored) ids as unclaimed and clean them up.""" + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("outer-trace-finish-reason-del") session_id_var.set("outer-session-finish-reason-del") try: @@ -4338,10 +4343,11 @@ def test_stream_wrapper_del_cleans_up_after_synthesized_finish_reason_chunk(): @pytest.mark.asyncio -async def test_stream_wrapper_anext_keeps_context_active_through_synthesized_finish_reason_chunk(): +async def test_stream_wrapper_anext_keeps_context_active_through_synthesized_finish_reason_chunk(monkeypatch): """Async sibling of test_stream_wrapper_next_keeps_context_active_through_synthesized_finish_reason_chunk - _finalize_completed_stream()'s else branch must not restore before returning the synthesized chunk either.""" + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("outer-trace-anext-finish-reason") session_id_var.set("outer-session-anext-finish-reason") try: @@ -4394,6 +4400,7 @@ async def test_stream_wrapper_anext_max_duration_timeout_restores_consumer_corre path as every other failure so the consumer's outer correlation context gets restored - calling the check before entering __anext__()'s try block would let the Timeout bypass that restoration entirely.""" + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) monkeypatch.setattr(litellm.constants, "LITELLM_MAX_STREAMING_DURATION_SECONDS", 1) trace_id_var.set("outer-trace-max-duration") session_id_var.set("outer-session-max-duration") @@ -4434,12 +4441,13 @@ async def test_stream_wrapper_anext_max_duration_timeout_restores_consumer_corre @pytest.mark.asyncio -async def test_stream_wrapper_aclose_restores_consumer_correlation_context(): +async def test_stream_wrapper_aclose_restores_consumer_correlation_context(monkeypatch): """Explicit early termination (aclose(), e.g. on client disconnect or a router fallback aborting an in-progress stream) must restore the caller's correlation context too - not just __del__'s best-effort GC-timed fallback, since aclose() is normally called deterministically by the consumer/ framework, unlike __del__.""" + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("outer-trace-aclose") session_id_var.set("outer-session-aclose") try: @@ -4481,6 +4489,7 @@ async def test_stream_wrapper_aclose_keeps_context_active_through_close_failure_ branch logs a debug diagnostic. That log line must still carry the closing stream's own trace_id/session_id - the outer context must not be restored until after the close attempt (and its diagnostic) completes.""" + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("outer-trace-close-fail") session_id_var.set("outer-session-close-fail") try: @@ -4541,6 +4550,7 @@ def test_handle_stream_fallback_error_restores_context_only_after_exception_mapp mapping. The consumer's outer context must not be restored until that mapping call returns, or the diagnostic log line would carry the outer (or empty) trace_id/session_id instead of the failing stream's own.""" + monkeypatch.setattr(litellm, "request_correlation_in_logs", True) trace_id_var.set("outer-trace-fallback") session_id_var.set("outer-session-fallback") try: From 1661e72c2c53a63c010685c04354bcaabcd12a97 Mon Sep 17 00:00:00 2001 From: mateo Date: Mon, 14 Sep 2026 13:21:46 +0000 Subject: [PATCH 3/4] fix(registry): drop retired friendliai llama-3.1 serverless models Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> --- ...odel_prices_and_context_window_backup.json | 28 ------------------- model_prices_and_context_window.json | 28 ------------------- 2 files changed, 56 deletions(-) diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json index 86605d77aab..fa67d8de443 100644 --- a/litellm/model_prices_and_context_window_backup.json +++ b/litellm/model_prices_and_context_window_backup.json @@ -23164,34 +23164,6 @@ "output_cost_per_token": 0.0, "source": "https://fireworks.ai/pricing" }, - "friendliai/meta-llama-3.1-70b-instruct": { - "input_cost_per_token": 6e-07, - "litellm_provider": "friendliai", - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 6e-07, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "friendliai/meta-llama-3.1-8b-instruct": { - "input_cost_per_token": 1e-07, - "litellm_provider": "friendliai", - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1e-07, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, "friendliai/zai-org/GLM-5.3-Flash": { "litellm_provider": "friendliai", "max_input_tokens": 1048576, diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index 86605d77aab..fa67d8de443 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -23164,34 +23164,6 @@ "output_cost_per_token": 0.0, "source": "https://fireworks.ai/pricing" }, - "friendliai/meta-llama-3.1-70b-instruct": { - "input_cost_per_token": 6e-07, - "litellm_provider": "friendliai", - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 6e-07, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "friendliai/meta-llama-3.1-8b-instruct": { - "input_cost_per_token": 1e-07, - "litellm_provider": "friendliai", - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1e-07, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, "friendliai/zai-org/GLM-5.3-Flash": { "litellm_provider": "friendliai", "max_input_tokens": 1048576, From 32699a0be91c66c907e93c61eb2969d3c52dfee2 Mon Sep 17 00:00:00 2001 From: "devin-ai-integration[bot]" <158243242+devin-ai-integration[bot]@users.noreply.github.com> Date: Sun, 13 Sep 2026 21:46:25 -0700 Subject: [PATCH 4/4] build(deps): re-suppress GHSA-h7x2-h6g9-p789 in osv-scan, mlflow still has no fixed release (#41036) Co-authored-by: mateo Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> --- osv-scanner.toml | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/osv-scanner.toml b/osv-scanner.toml index 3e070fc8cf7..482254d4da6 100644 --- a/osv-scanner.toml +++ b/osv-scanner.toml @@ -5,5 +5,5 @@ reason = "diskcache has no fixed release published; remove this entry once one e [[IgnoredVulns]] id = "GHSA-h7x2-h6g9-p789" -ignoreUntil = 2026-09-14 -reason = "mlflow has no fixed release published; remove this entry once one exists" +ignoreUntil = 2026-10-14 +reason = "mlflow has no fixed release published (3.16.0, 2026-09-04, and master still store gateway secret api_base unvalidated); remove this entry once one exists"