scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-08-28 05:25:59 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 36

fix(model_prices): correct gemini-3.5-flash-lite flex cache-read pricing

Browse source
This commit is contained in:
mateo-berri 2026-08-26 16:25:15 -07:00
parent f7220556e1
commit 0243c5dee4
3 changed files with 44 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
litellm/model_prices_and_context_window_backup.json
View file

@ -20147,7 +20147,7 @@
"gemini-3.5-flash-lite": {
"deprecation_date": "2027-07-21",
"cache_read_input_token_cost": 3e-08,
"cache_read_input_token_cost_flex": 2e-08,
"cache_read_input_token_cost_flex": 1.5e-08,
"cache_read_input_token_cost_priority": 5e-08,
"input_cost_per_token": 3e-07,
"input_cost_per_token_batches": 1.5e-07,
@ -22488,7 +22488,7 @@
},
"gemini/gemini-3.5-flash-lite": {
"cache_read_input_token_cost": 3e-08,
"cache_read_input_token_cost_flex": 2e-08,
"cache_read_input_token_cost_flex": 1.5e-08,
"cache_read_input_token_cost_priority": 5e-08,
"input_cost_per_token": 3e-07,
"input_cost_per_token_batches": 1.5e-07,
@ -41987,7 +41987,7 @@
"vertex_ai/gemini-3.5-flash-lite": {
"deprecation_date": "2027-07-21",
"cache_read_input_token_cost": 3e-08,
"cache_read_input_token_cost_flex": 2e-08,
"cache_read_input_token_cost_flex": 1.5e-08,
"cache_read_input_token_cost_priority": 5e-08,
"input_cost_per_token": 3e-07,
"input_cost_per_token_batches": 1.5e-07,

6
model_prices_and_context_window.json
View file

@ -20147,7 +20147,7 @@
"gemini-3.5-flash-lite": {
"deprecation_date": "2027-07-21",
"cache_read_input_token_cost": 3e-08,
"cache_read_input_token_cost_flex": 2e-08,
"cache_read_input_token_cost_flex": 1.5e-08,
"cache_read_input_token_cost_priority": 5e-08,
"input_cost_per_token": 3e-07,
"input_cost_per_token_batches": 1.5e-07,
@ -22488,7 +22488,7 @@
},
"gemini/gemini-3.5-flash-lite": {
"cache_read_input_token_cost": 3e-08,
"cache_read_input_token_cost_flex": 2e-08,
"cache_read_input_token_cost_flex": 1.5e-08,
"cache_read_input_token_cost_priority": 5e-08,
"input_cost_per_token": 3e-07,
"input_cost_per_token_batches": 1.5e-07,
@ -41987,7 +41987,7 @@
"vertex_ai/gemini-3.5-flash-lite": {
"deprecation_date": "2027-07-21",
"cache_read_input_token_cost": 3e-08,
"cache_read_input_token_cost_flex": 2e-08,
"cache_read_input_token_cost_flex": 1.5e-08,
"cache_read_input_token_cost_priority": 5e-08,
"input_cost_per_token": 3e-07,
"input_cost_per_token_batches": 1.5e-07,

38
tests/test_litellm/litellm_core_utils/llm_cost_calc/test_llm_cost_calc_utils.py
View file

@ -3377,6 +3377,44 @@ def test_generic_cost_per_token_gemini_35_flash_lite(_local_model_cost_map):
assert completion_cost == pytest.approx(0.00125)
GEMINI_35_FLASH_LITE_SERVICE_TIER_PRICING = [
(None, 3e-07, 2.5e-06, 3e-08),
("flex", 1.5e-07, 1.25e-06, 1.5e-08),
("priority", 5.4e-07, 4.5e-06, 5e-08),
]
@pytest.mark.parametrize(
"service_tier,input_rate,output_rate,cache_read_rate", GEMINI_35_FLASH_LITE_SERVICE_TIER_PRICING
)
@pytest.mark.parametrize(
"model",
["gemini-3.5-flash-lite", "gemini/gemini-3.5-flash-lite", "vertex_ai/gemini-3.5-flash-lite"],
)
def test_gemini_35_flash_lite_service_tier_pricing(
model, service_tier, input_rate, output_rate, cache_read_rate, _local_model_cost_map
):
"""Regression: Vertex publishes flash-lite Flex/Batch context caching at $0.015/M
(1.5e-08/token), so flex cache reads must not be billed at the 2e-08 rate the map
used to carry."""
usage = Usage(
prompt_tokens=1_000,
completion_tokens=500,
total_tokens=1_500,
prompt_tokens_details=PromptTokensDetailsWrapper(cached_tokens=200, text_tokens=800),
)
prompt_cost, completion_cost = generic_cost_per_token(
model=model.split("/")[-1],
usage=usage,
custom_llm_provider=model.split("/")[0] if "/" in model else "gemini",
service_tier=service_tier,
)
assert prompt_cost == pytest.approx(800 * input_rate + 200 * cache_read_rate, rel=1e-9)
assert completion_cost == pytest.approx(500 * output_rate, rel=1e-9)
@pytest.mark.parametrize(
"service_tier,input_rate,cache_read_rate,cache_write_rate,output_rate",
[