From 3202963f25482bb3a2fe312f93f1f12291f7d60e Mon Sep 17 00:00:00 2001 From: mateo-berri <277851410+mateo-berri@users.noreply.github.com> Date: Fri, 4 Sep 2026 16:57:40 -0700 Subject: [PATCH] feat(cost-map): add azure/gpt-6-astra and azure/us/gpt-6-astra Foundry pricing --- ...odel_prices_and_context_window_backup.json | 94 +++++++++++++++++++ model_prices_and_context_window.json | 94 +++++++++++++++++++ .../llm_cost_calc/test_llm_cost_calc_utils.py | 43 +++++++++ .../test_reasoning_effort_capability.py | 16 ++++ 4 files changed, 247 insertions(+) diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json index 7f5038e3073..d8b7287eba7 100644 --- a/litellm/model_prices_and_context_window_backup.json +++ b/litellm/model_prices_and_context_window_backup.json @@ -7157,6 +7157,53 @@ "supports_xhigh_reasoning_effort": true, "supports_minimal_reasoning_effort": false }, + "azure/gpt-6-astra": { + "cache_creation_input_token_cost": 1.25e-05, + "cache_creation_input_token_cost_above_272k_tokens": 2.5e-05, + "cache_read_input_token_cost": 1e-06, + "cache_read_input_token_cost_above_272k_tokens": 2e-06, + "input_cost_per_token": 1e-05, + "input_cost_per_token_above_272k_tokens": 2e-05, + "litellm_provider": "azure", + "max_input_tokens": 922000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 5e-05, + "output_cost_per_token_above_272k_tokens": 7.5e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_computer_use": true, + "supports_function_calling": true, + "supports_max_reasoning_effort": true, + "supports_minimal_reasoning_effort": false, + "supports_native_streaming": true, + "supports_none_reasoning_effort": false, + "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, + "supports_xhigh_reasoning_effort": 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, @@ -7376,6 +7423,53 @@ "supports_xhigh_reasoning_effort": true, "supports_minimal_reasoning_effort": false }, + "azure/us/gpt-6-astra": { + "cache_creation_input_token_cost": 1.375e-05, + "cache_creation_input_token_cost_above_272k_tokens": 2.75e-05, + "cache_read_input_token_cost": 1.1e-06, + "cache_read_input_token_cost_above_272k_tokens": 2.2e-06, + "input_cost_per_token": 1.1e-05, + "input_cost_per_token_above_272k_tokens": 2.2e-05, + "litellm_provider": "azure", + "max_input_tokens": 922000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 5.5e-05, + "output_cost_per_token_above_272k_tokens": 8.25e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_computer_use": true, + "supports_function_calling": true, + "supports_max_reasoning_effort": true, + "supports_minimal_reasoning_effort": false, + "supports_native_streaming": true, + "supports_none_reasoning_effort": false, + "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, + "supports_xhigh_reasoning_effort": 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 7f5038e3073..d8b7287eba7 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -7157,6 +7157,53 @@ "supports_xhigh_reasoning_effort": true, "supports_minimal_reasoning_effort": false }, + "azure/gpt-6-astra": { + "cache_creation_input_token_cost": 1.25e-05, + "cache_creation_input_token_cost_above_272k_tokens": 2.5e-05, + "cache_read_input_token_cost": 1e-06, + "cache_read_input_token_cost_above_272k_tokens": 2e-06, + "input_cost_per_token": 1e-05, + "input_cost_per_token_above_272k_tokens": 2e-05, + "litellm_provider": "azure", + "max_input_tokens": 922000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 5e-05, + "output_cost_per_token_above_272k_tokens": 7.5e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_computer_use": true, + "supports_function_calling": true, + "supports_max_reasoning_effort": true, + "supports_minimal_reasoning_effort": false, + "supports_native_streaming": true, + "supports_none_reasoning_effort": false, + "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, + "supports_xhigh_reasoning_effort": 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, @@ -7376,6 +7423,53 @@ "supports_xhigh_reasoning_effort": true, "supports_minimal_reasoning_effort": false }, + "azure/us/gpt-6-astra": { + "cache_creation_input_token_cost": 1.375e-05, + "cache_creation_input_token_cost_above_272k_tokens": 2.75e-05, + "cache_read_input_token_cost": 1.1e-06, + "cache_read_input_token_cost_above_272k_tokens": 2.2e-06, + "input_cost_per_token": 1.1e-05, + "input_cost_per_token_above_272k_tokens": 2.2e-05, + "litellm_provider": "azure", + "max_input_tokens": 922000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 5.5e-05, + "output_cost_per_token_above_272k_tokens": 8.25e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_computer_use": true, + "supports_function_calling": true, + "supports_max_reasoning_effort": true, + "supports_minimal_reasoning_effort": false, + "supports_native_streaming": true, + "supports_none_reasoning_effort": false, + "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, + "supports_xhigh_reasoning_effort": 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 0f8084643ea..df680b7cb0e 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 @@ -2008,6 +2008,49 @@ def test_generic_cost_per_token_azure_gpt56(_local_model_cost_map, assert round(completion_cost, 10) == round(output_cost * completion_tokens, 10) +@pytest.mark.parametrize("model,zone_multiplier", [("azure/gpt-6-astra", 1.0), ("azure/us/gpt-6-astra", 1.1)]) +@pytest.mark.parametrize( + "prompt_tokens,input_side_multiplier,output_multiplier", + [(100000, 1.0, 1.0), (300000, 2.0, 1.5)], +) +def test_generic_cost_per_token_azure_gpt_6_astra_foundry_price_sheet( + _local_model_cost_map, + model, + zone_multiplier, + prompt_tokens, + input_side_multiplier, + output_multiplier, +): + """Microsoft Foundry sells gpt-6-astra at the OpenAI rates: $10 input, $1 cache read, $12.50 cache write, + $50 output per 1M tokens on Standard Global, with the input side doubling and output 1.5x above 272K + prompt tokens. Standard US Data Zone carries the usual 10% uplift on every rate. + """ + cached_tokens = 50000 + cache_write_tokens = 40000 + text_tokens = prompt_tokens - cached_tokens - cache_write_tokens + 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, cache_write_tokens=cache_write_tokens + ), + ) + + prompt_cost, completion_cost = generic_cost_per_token( + model=model, + usage=usage, + custom_llm_provider="azure", + ) + + input_side = zone_multiplier * input_side_multiplier + assert prompt_cost == pytest.approx( + input_side * (text_tokens * 1e-5 + cached_tokens * 1e-6 + cache_write_tokens * 1.25e-5) + ) + assert completion_cost == pytest.approx(zone_multiplier * output_multiplier * completion_tokens * 5e-5) + + @pytest.mark.parametrize( "model,expected_none,expected_xhigh,expected_minimal", [ diff --git a/tests/test_litellm/router_utils/test_reasoning_effort_capability.py b/tests/test_litellm/router_utils/test_reasoning_effort_capability.py index 7b2e45ab3ed..504e87fb231 100644 --- a/tests/test_litellm/router_utils/test_reasoning_effort_capability.py +++ b/tests/test_litellm/router_utils/test_reasoning_effort_capability.py @@ -388,3 +388,19 @@ class TestGpt6AstraAdvertisesItsDocumentedLevels: "xhigh", "max", ) + + @pytest.mark.parametrize("model", ["azure/gpt-6-astra", "azure/us/gpt-6-astra"]) + def test_a_foundry_deployment_advertises_the_same_levels(self, local_model_cost_map, model): + """Microsoft Foundry serves the same model, so an Azure deployment must offer low + through max and never none, exactly like the OpenAI entry.""" + from litellm.utils import _get_model_info_helper + + model_info = dict(_get_model_info_helper(model=model, custom_llm_provider="azure")) + + assert resolve_supported_reasoning_efforts(model_info, deployment_is_mapped=True) == ( + "low", + "medium", + "high", + "xhigh", + "max", + )