From 91c7f758640b5d60c3fcfd9e51bcd1770a74fd25 Mon Sep 17 00:00:00 2001 From: kerry Date: Wed, 16 Sep 2026 01:44:30 +0000 Subject: [PATCH] docs(fireworks-ai): describe the cache-read fallback without asserting provider billing Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> --- litellm/llms/fireworks_ai/cost_calculator.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/litellm/llms/fireworks_ai/cost_calculator.py b/litellm/llms/fireworks_ai/cost_calculator.py index a92e5208471..1795a700d25 100644 --- a/litellm/llms/fireworks_ai/cost_calculator.py +++ b/litellm/llms/fireworks_ai/cost_calculator.py @@ -68,8 +68,8 @@ def _resolve_model_info(model: str) -> ModelInfo: def _with_cache_read_fallback(model_info: ModelInfo) -> ModelInfo: - """Most fireworks_ai price-map entries publish no cache-read rate though the provider bills - cached reads at the input rate; the shared map is never mutated, so a copy carries the fallback.""" + """Entries without a cache-read rate keep the previous calculator's input-rate fallback for cached + reads (LIT-7845 tracks the documented discount); the shared map is never mutated, so a copy carries it.""" input_rate: Final = model_info.get("input_cost_per_token") if model_info.get("cache_read_input_token_cost") is not None or input_rate is None: return model_info