scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-09-14 23:21:35 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 102

fix(test): use model with cache pricing so assertion actually executes

Browse source 
Switch test from `fireworks_ai/llama-v3p3-70b-instruct` (no
cache_read_input_token_cost) to `fireworks_ai/kimi-k2p5` (has
cache pricing at 1e-07 vs input 6e-07). Remove the conditional
guard so the assertion always runs.

Addresses Greptile review feedback on #24860.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Signed-off-by: GopalGB <67310594+GopalGB@users.noreply.github.com>
This commit is contained in:
GopalGB 2026-05-07 20:23:04 +05:30
parent 73112c87ac
commit a7a0c328d4
2 changed files with 9 additions and 15 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
litellm/llms/fireworks_ai/cost_calculator.py
View file

@ -91,9 +91,7 @@ def cost_per_token(model: str, usage: Usage) -> Tuple[float, float]:
)
## ADJUST FOR CACHE CREATION TOKENS
cache_creation_input_tokens = (
usage.get("cache_creation_input_tokens", None) or 0
)
cache_creation_input_tokens = usage.get("cache_creation_input_tokens", None) or 0
cache_creation_cost = model_info.get("cache_creation_input_token_cost", None)
if cache_creation_input_tokens > 0 and cache_creation_cost is not None:
prompt_cost += cache_creation_input_tokens * (

20
tests/local_testing/test_completion_cost.py
View file

@ -1221,27 +1221,23 @@ def test_fireworks_ai_cache_token_pricing():
os.environ["LITELLM_LOCAL_MODEL_COST_MAP"] = "True"
litellm.model_cost = litellm.get_model_cost_map(url="")
# Use kimi-k2p5 which has cache_read_input_token_cost in the pricing config
prompt_cost_cached, completion_cost_cached = cost_per_token(
model="fireworks_ai/llama-v3p3-70b-instruct", usage=usage_with_cache
model="fireworks_ai/kimi-k2p5", usage=usage_with_cache
)
prompt_cost_no_cache, completion_cost_no_cache = cost_per_token(
model="fireworks_ai/llama-v3p3-70b-instruct", usage=usage_no_cache
model="fireworks_ai/kimi-k2p5", usage=usage_no_cache
)
# Completion cost should be the same regardless of cache
assert completion_cost_cached == completion_cost_no_cache
# If the model has cache pricing, the prompt cost with cache should differ
# from the prompt cost without cache (cache read rate is cheaper)
model_info = litellm.get_model_info(
model="fireworks_ai/llama-v3p3-70b-instruct",
custom_llm_provider="fireworks_ai",
# kimi-k2p5 has cache_read_input_token_cost (1e-07) < input_cost_per_token (6e-07),
# so prompt cost with 800 cache-read tokens must be cheaper
assert prompt_cost_cached < prompt_cost_no_cache, (
"Prompt cost with 800 cache-read tokens should be less than "
"full-price for the same total prompt tokens"
)
if model_info.get("cache_read_input_token_cost") is not None:
assert prompt_cost_cached < prompt_cost_no_cache, (
"Prompt cost with 800 cache-read tokens should be less than "
"full-price for the same total prompt tokens"
)
def test_cost_azure_openai_prompt_caching():