From a17c8b8ef9a081badfd39ed9ee443f3e52f1c661 Mon Sep 17 00:00:00 2001 From: Zero Clover Date: Sat, 13 Jun 2026 14:49:08 +0800 Subject: [PATCH] fix(pricing): correct grok-4.20 series rates and context window The built-in grok-4.20 entries billed input at $2/M and output at $6/M with a 2M context window. x.ai documents $1.25/M input, $2.50/M output, $0.20/M cached input and a 1M context window for the non-reasoning, reasoning and multi-agent variants, so every request was overbilled and the advertised context window was double the real one Corrects input_cost_per_token, output_cost_per_token, max_input_tokens, max_output_tokens and max_tokens for the four xai/grok-4.20-* keys and the two vertex_ai/xai/grok-4.20-* keys (which mirror x.ai pricing), in both model_prices_and_context_window.json and the bundled backup. The cached input rate was already correct and is left untouched Sources: https://docs.x.ai/developers/models/grok-4.20-0309-non-reasoning https://docs.x.ai/developers/models/grok-4.20-0309-reasoning https://docs.x.ai/developers/models/grok-4.20-multi-agent-0309 --- ...odel_prices_and_context_window_backup.json | 60 +++++++++---------- model_prices_and_context_window.json | 60 +++++++++---------- 2 files changed, 60 insertions(+), 60 deletions(-) diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json index 76a7c0640af..50bfffb8c88 100644 --- a/litellm/model_prices_and_context_window_backup.json +++ b/litellm/model_prices_and_context_window_backup.json @@ -35504,13 +35504,13 @@ }, "vertex_ai/xai/grok-4.20-non-reasoning": { "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 2e-06, + "input_cost_per_token": 1.25e-06, "litellm_provider": "vertex_ai", - "max_input_tokens": 2000000, - "max_output_tokens": 2000000, - "max_tokens": 2000000, + "max_input_tokens": 1000000, + "max_output_tokens": 1000000, + "max_tokens": 1000000, "mode": "chat", - "output_cost_per_token": 6e-06, + "output_cost_per_token": 2.5e-06, "source": "https://docs.x.ai/docs/models (Vertex AI Model Garden)", "supports_function_calling": true, "supports_response_schema": true, @@ -35520,13 +35520,13 @@ }, "vertex_ai/xai/grok-4.20-reasoning": { "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 2e-06, + "input_cost_per_token": 1.25e-06, "litellm_provider": "vertex_ai", - "max_input_tokens": 2000000, - "max_output_tokens": 2000000, - "max_tokens": 2000000, + "max_input_tokens": 1000000, + "max_output_tokens": 1000000, + "max_tokens": 1000000, "mode": "chat", - "output_cost_per_token": 6e-06, + "output_cost_per_token": 2.5e-06, "source": "https://docs.x.ai/docs/models (Vertex AI Model Garden)", "supports_function_calling": true, "supports_reasoning": true, @@ -36840,13 +36840,13 @@ }, "xai/grok-4.20-multi-agent-beta-0309": { "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 2e-06, + "input_cost_per_token": 1.25e-06, "litellm_provider": "xai", - "max_input_tokens": 2000000, - "max_output_tokens": 2000000, - "max_tokens": 2000000, + "max_input_tokens": 1000000, + "max_output_tokens": 1000000, + "max_tokens": 1000000, "mode": "chat", - "output_cost_per_token": 6e-06, + "output_cost_per_token": 2.5e-06, "source": "https://docs.x.ai/docs/models", "supports_function_calling": true, "supports_prompt_caching": true, @@ -36857,13 +36857,13 @@ }, "xai/grok-4.20-beta-0309-reasoning": { "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 2e-06, + "input_cost_per_token": 1.25e-06, "litellm_provider": "xai", - "max_input_tokens": 2000000, - "max_output_tokens": 2000000, - "max_tokens": 2000000, + "max_input_tokens": 1000000, + "max_output_tokens": 1000000, + "max_tokens": 1000000, "mode": "chat", - "output_cost_per_token": 6e-06, + "output_cost_per_token": 2.5e-06, "source": "https://docs.x.ai/docs/models", "supports_function_calling": true, "supports_prompt_caching": true, @@ -36874,13 +36874,13 @@ }, "xai/grok-4.20-0309-reasoning": { "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 2e-06, + "input_cost_per_token": 1.25e-06, "litellm_provider": "xai", - "max_input_tokens": 2000000, - "max_output_tokens": 2000000, - "max_tokens": 2000000, + "max_input_tokens": 1000000, + "max_output_tokens": 1000000, + "max_tokens": 1000000, "mode": "chat", - "output_cost_per_token": 6e-06, + "output_cost_per_token": 2.5e-06, "source": "https://docs.x.ai/docs/models", "supports_function_calling": true, "supports_reasoning": true, @@ -36890,13 +36890,13 @@ }, "xai/grok-4.20-beta-0309-non-reasoning": { "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 2e-06, + "input_cost_per_token": 1.25e-06, "litellm_provider": "xai", - "max_input_tokens": 2000000, - "max_output_tokens": 2000000, - "max_tokens": 2000000, + "max_input_tokens": 1000000, + "max_output_tokens": 1000000, + "max_tokens": 1000000, "mode": "chat", - "output_cost_per_token": 6e-06, + "output_cost_per_token": 2.5e-06, "source": "https://docs.x.ai/docs/models", "supports_function_calling": true, "supports_prompt_caching": true, diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index b181df94131..873de3ef98e 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -35544,13 +35544,13 @@ }, "vertex_ai/xai/grok-4.20-non-reasoning": { "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 2e-06, + "input_cost_per_token": 1.25e-06, "litellm_provider": "vertex_ai", - "max_input_tokens": 2000000, - "max_output_tokens": 2000000, - "max_tokens": 2000000, + "max_input_tokens": 1000000, + "max_output_tokens": 1000000, + "max_tokens": 1000000, "mode": "chat", - "output_cost_per_token": 6e-06, + "output_cost_per_token": 2.5e-06, "source": "https://docs.x.ai/docs/models (Vertex AI Model Garden)", "supports_function_calling": true, "supports_response_schema": true, @@ -35560,13 +35560,13 @@ }, "vertex_ai/xai/grok-4.20-reasoning": { "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 2e-06, + "input_cost_per_token": 1.25e-06, "litellm_provider": "vertex_ai", - "max_input_tokens": 2000000, - "max_output_tokens": 2000000, - "max_tokens": 2000000, + "max_input_tokens": 1000000, + "max_output_tokens": 1000000, + "max_tokens": 1000000, "mode": "chat", - "output_cost_per_token": 6e-06, + "output_cost_per_token": 2.5e-06, "source": "https://docs.x.ai/docs/models (Vertex AI Model Garden)", "supports_function_calling": true, "supports_reasoning": true, @@ -36870,13 +36870,13 @@ }, "xai/grok-4.20-multi-agent-beta-0309": { "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 2e-06, + "input_cost_per_token": 1.25e-06, "litellm_provider": "xai", - "max_input_tokens": 2000000, - "max_output_tokens": 2000000, - "max_tokens": 2000000, + "max_input_tokens": 1000000, + "max_output_tokens": 1000000, + "max_tokens": 1000000, "mode": "chat", - "output_cost_per_token": 6e-06, + "output_cost_per_token": 2.5e-06, "source": "https://docs.x.ai/docs/models", "supports_function_calling": true, "supports_prompt_caching": true, @@ -36887,13 +36887,13 @@ }, "xai/grok-4.20-beta-0309-reasoning": { "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 2e-06, + "input_cost_per_token": 1.25e-06, "litellm_provider": "xai", - "max_input_tokens": 2000000, - "max_output_tokens": 2000000, - "max_tokens": 2000000, + "max_input_tokens": 1000000, + "max_output_tokens": 1000000, + "max_tokens": 1000000, "mode": "chat", - "output_cost_per_token": 6e-06, + "output_cost_per_token": 2.5e-06, "source": "https://docs.x.ai/docs/models", "supports_function_calling": true, "supports_prompt_caching": true, @@ -36904,13 +36904,13 @@ }, "xai/grok-4.20-0309-reasoning": { "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 2e-06, + "input_cost_per_token": 1.25e-06, "litellm_provider": "xai", - "max_input_tokens": 2000000, - "max_output_tokens": 2000000, - "max_tokens": 2000000, + "max_input_tokens": 1000000, + "max_output_tokens": 1000000, + "max_tokens": 1000000, "mode": "chat", - "output_cost_per_token": 6e-06, + "output_cost_per_token": 2.5e-06, "source": "https://docs.x.ai/docs/models", "supports_function_calling": true, "supports_reasoning": true, @@ -36920,13 +36920,13 @@ }, "xai/grok-4.20-beta-0309-non-reasoning": { "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 2e-06, + "input_cost_per_token": 1.25e-06, "litellm_provider": "xai", - "max_input_tokens": 2000000, - "max_output_tokens": 2000000, - "max_tokens": 2000000, + "max_input_tokens": 1000000, + "max_output_tokens": 1000000, + "max_tokens": 1000000, "mode": "chat", - "output_cost_per_token": 6e-06, + "output_cost_per_token": 2.5e-06, "source": "https://docs.x.ai/docs/models", "supports_function_calling": true, "supports_prompt_caching": true,