scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-10-08 03:08:45 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 114

(fix): token_count calculation during inline data sent during multimodal inputs

Browse source
This commit is contained in:
Praveen11558 2026-03-10 15:23:52 +05:30 • committed by GitHub
parent dd3bcd3421
commit 69479ab69b
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 33 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

33
tests/test_litellm/litellm_core_utils/llm_cost_calc/test_llm_cost_calc_utils.py
View file

@ -126,7 +126,40 @@ def test_reasoning_tokens_gemini():
10,
)
def test_inline_pdf_tokens_costed_when_text_tokens_less_than_prompt_tokens():
usage = Usage(
completion_tokens=100,
prompt_tokens=5200,
total_tokens=5300,
prompt_tokens_details=PromptTokensDetailsWrapper(
audio_tokens=None, cached_tokens=None, text_tokens=12, image_tokens=None
),
)
response = ModelResponse(
usage= usage,
model = "gemini-2.0-flash-001",
)
cost = response_cost_calculator(
reponse_object= response,
model = "gemini-2.0-flash-001",
custom_llm_provider = "vertex_ai",
call_type = "acompletion",
optional_params= {},
)
model_info = litellm.model_cost["gemini-2.0-flash-001"]
input_cost_per_token = model_info["input_cost_per_token"]
output_cost_per_token = model_info["output_cost_per_token"]
expected_cost = (
5200* input_cost_per_token + 100 * output_cost_per_token
)
assert cost == pytest.approx(expected_cost), (
f"Expected cost={expected_cost} (all 5200 prompt tokens costed), "
f"got cost={cost}. PDF tokens are likely not being costed."
)
def test_reasoning_tokens_gemini_3_1_flash_lite():
"""Test cost calculation for gemini-3.1-flash-lite-preview with reasoning tokens"""
model = "gemini-3.1-flash-lite-preview"