From a1e6b8650376f450a46708b3339044052c764c99 Mon Sep 17 00:00:00 2001 From: Sushant Lokhande Date: Sun, 5 Jul 2026 02:57:12 -0700 Subject: [PATCH] fix: correct max_output_tokens for Claude Haiku 4.5 in vertex_ai and openrouter entries vertex_ai/claude-haiku-4-5 and vertex_ai/claude-haiku-4-5@20251001 said 8192 (stale carry-over from Claude 3.5 Haiku); Google's Vertex AI docs list 64,000 max output tokens for Claude Haiku 4.5. openrouter/anthropic/claude-haiku-4.5 said 200000 (the context window pasted into the output fields); OpenRouter's /api/v1/models reports top_provider.max_completion_tokens: 64000. Backup catalog updated to match. Fixes #32184 --- litellm/model_prices_and_context_window_backup.json | 12 ++++++------ model_prices_and_context_window.json | 12 ++++++------ 2 files changed, 12 insertions(+), 12 deletions(-) diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json index 26d54c05ffb..18f7fe8622d 100644 --- a/litellm/model_prices_and_context_window_backup.json +++ b/litellm/model_prices_and_context_window_backup.json @@ -28548,8 +28548,8 @@ "input_cost_per_token": 1e-06, "litellm_provider": "openrouter", "max_input_tokens": 200000, - "max_output_tokens": 200000, - "max_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, "mode": "chat", "output_cost_per_token": 5e-06, "supports_assistant_prefill": true, @@ -34688,8 +34688,8 @@ "input_cost_per_token": 1e-06, "litellm_provider": "vertex_ai-anthropic_models", "max_input_tokens": 200000, - "max_output_tokens": 8192, - "max_tokens": 8192, + "max_output_tokens": 64000, + "max_tokens": 64000, "mode": "chat", "output_cost_per_token": 5e-06, "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude/haiku-4-5", @@ -34710,8 +34710,8 @@ "input_cost_per_token": 1e-06, "litellm_provider": "vertex_ai-anthropic_models", "max_input_tokens": 200000, - "max_output_tokens": 8192, - "max_tokens": 8192, + "max_output_tokens": 64000, + "max_tokens": 64000, "mode": "chat", "output_cost_per_token": 5e-06, "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude/haiku-4-5", diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index a3dcdaed3f7..6ca25bc5819 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -28706,8 +28706,8 @@ "input_cost_per_token": 1e-06, "litellm_provider": "openrouter", "max_input_tokens": 200000, - "max_output_tokens": 200000, - "max_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, "mode": "chat", "output_cost_per_token": 5e-06, "supports_assistant_prefill": true, @@ -34862,8 +34862,8 @@ "input_cost_per_token": 1e-06, "litellm_provider": "vertex_ai-anthropic_models", "max_input_tokens": 200000, - "max_output_tokens": 8192, - "max_tokens": 8192, + "max_output_tokens": 64000, + "max_tokens": 64000, "mode": "chat", "output_cost_per_token": 5e-06, "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude/haiku-4-5", @@ -34884,8 +34884,8 @@ "input_cost_per_token": 1e-06, "litellm_provider": "vertex_ai-anthropic_models", "max_input_tokens": 200000, - "max_output_tokens": 8192, - "max_tokens": 8192, + "max_output_tokens": 64000, + "max_tokens": 64000, "mode": "chat", "output_cost_per_token": 5e-06, "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude/haiku-4-5",