From 54868aae77d731a232019cab64bcd1a1650f87ca Mon Sep 17 00:00:00 2001 From: Bartok9 Date: Mon, 20 Jul 2026 19:10:51 -0400 Subject: [PATCH] fix(pricing): correct azure_ai Llama-4-Maverick token costs salvage: rebuild of @Kropiunig work on pricing for azure_ai/Llama-4-Maverick-17B-128E-Instruct-FP8 (see prior discussion / open attempt). Input/output were inverted vs Azure retail global standard ($0.25/M in, $1.00/M out). Sync primary map + backup twin. Credit: @Kropiunig AI-assisted; human-reviewed. --- litellm/model_prices_and_context_window_backup.json | 4 ++-- model_prices_and_context_window.json | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json index e5afc81b641..46035a59223 100644 --- a/litellm/model_prices_and_context_window_backup.json +++ b/litellm/model_prices_and_context_window_backup.json @@ -8441,13 +8441,13 @@ "supports_tool_choice": true }, "azure_ai/Llama-4-Maverick-17B-128E-Instruct-FP8": { - "input_cost_per_token": 1.41e-06, + "input_cost_per_token": 2.5e-07, "litellm_provider": "azure_ai", "max_input_tokens": 1000000, "max_output_tokens": 16384, "max_tokens": 16384, "mode": "chat", - "output_cost_per_token": 3.5e-07, + "output_cost_per_token": 1e-06, "source": "https://azure.microsoft.com/en-us/blog/introducing-the-llama-4-herd-in-azure-ai-foundry-and-azure-databricks/", "supports_function_calling": true, "supports_tool_choice": true, diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index 5cf99ba8bac..4a10ea3d5e0 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -8441,13 +8441,13 @@ "supports_tool_choice": true }, "azure_ai/Llama-4-Maverick-17B-128E-Instruct-FP8": { - "input_cost_per_token": 1.41e-06, + "input_cost_per_token": 2.5e-07, "litellm_provider": "azure_ai", "max_input_tokens": 1000000, "max_output_tokens": 16384, "max_tokens": 16384, "mode": "chat", - "output_cost_per_token": 3.5e-07, + "output_cost_per_token": 1e-06, "source": "https://azure.microsoft.com/en-us/blog/introducing-the-llama-4-herd-in-azure-ai-foundry-and-azure-databricks/", "supports_function_calling": true, "supports_tool_choice": true,