scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-11-14 09:03:16 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 214

test(cost): add gpt-6-luna 12h cache write, gemini above-200k flex and vertex tier cost cases

Browse source
This commit is contained in:
devin-ai-integration[bot] 2026-11-09 15:42:25 +00:00 • committed by GitHub
parent 0854b1b940
commit e74676961b
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
5 changed files with 97 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

16
tests/integration/cost_calculation/cases/gemini.py
View file

@ -3480,6 +3480,21 @@ GEMINI_GEMINI_3_1_PRO_PREVIEW_TIERED_INPUT_ABOVE_200K: Final = _gemini_case(
),
usage_fields={"serviceTier": "standard"},
)
GEMINI_GEMINI_3_1_PRO_PREVIEW_TIERED_FLEX_ABOVE_200K: Final = _gemini_case(
GEMINI_GEMINI_3_1_PRO_TIERED_INPUT_ABOVE_200K,
"gemini-gemini-3.1-pro-preview-tiered_flex_above_200k",
"gemini/gemini-3.1-pro-preview",
"gemini-3.1-pro-preview",
ExactExpected(
spend=0.539734, input_cost=0.53248, output_cost=0.007254, prompt_tokens=204800, completion_tokens=620
),
usage_fields={"serviceTier": "flex"},
).model_copy(
update={
"known_bug": "gemini route bills the standard above-200k rate (1.079468) when the Gemini API reply marks flex only "
"through usageMetadata.serviceTier, so the above_200k_tokens_flex prices are never applied"
}
)
GEMINI_GEMINI_3_1_PRO_PREVIEW_TIERED_CACHE_READ_ABOVE_200K: Final = _gemini_case(
GEMINI_GEMINI_3_1_PRO_TIERED_CACHE_READ_ABOVE_200K,
"gemini-gemini-3.1-pro-preview-tiered_cache_read_above_200k",
@ -3572,6 +3587,7 @@ CASES: Final[tuple[CostTrackingTestCase, ...]] = (
GEMINI_GEMINI_3_1_PRO_PREVIEW_CACHE_READ,
GEMINI_GEMINI_3_1_PRO_PREVIEW_TIERED_INPUT_ABOVE_200K,
GEMINI_GEMINI_3_1_PRO_PREVIEW_TIERED_CACHE_READ_ABOVE_200K,
GEMINI_GEMINI_3_1_PRO_PREVIEW_TIERED_FLEX_ABOVE_200K,
GEMINI_GEMINI_3_1_PRO_PREVIEW_REASONING,
GEMINI_GEMINI_3_1_PRO_PREVIEW_TOOL_CALL,
GEMINI_GEMINI_3_1_PRO_PREVIEW_SERVICE_TIER_PRIORITY,

12
tests/integration/cost_calculation/cases/openai.py
View file

@ -7036,8 +7036,8 @@ GPT_5_6_RESPONSES_CACHE_WRITE_12H: Final = CostTrackingTestCase(
),
)
GPT_6_LUNA_RESPONSES_CACHE_WRITE_12H_NO_12H_RATE: Final = CostTrackingTestCase(
name="gpt-6-luna-responses_cache_write_12h_no_12h_rate",
GPT_6_LUNA_RESPONSES_CACHE_WRITE_12H: Final = CostTrackingTestCase(
name="gpt-6-luna-responses_cache_write_12h",
covers="quota_management.spend_tracking.cost_matrix.logs_cost",
model="gpt-6-luna",
endpoint="/v1/responses",
@ -7077,12 +7077,12 @@ GPT_6_LUNA_RESPONSES_CACHE_WRITE_12H_NO_12H_RATE: Final = CostTrackingTestCase(
},
),
expected=ExactExpected(
spend=0.0016078,
input_cost=0.0014178,
spend=0.0018578,
input_cost=0.0016678,
output_cost=0.00019,
prompt_tokens=12928,
completion_tokens=380,
cache_creation_cost=0.000625,
cache_creation_cost=0.000875,
),
)
@ -12224,7 +12224,7 @@ CASES: Final[tuple[CostTrackingTestCase, ...]] = (
OMNI_MODERATIONS_NEXT_LIST,
GPT_5_6_RESPONSES_CACHE_READ,
GPT_5_6_RESPONSES_CACHE_WRITE_12H,
GPT_6_LUNA_RESPONSES_CACHE_WRITE_12H_NO_12H_RATE,
GPT_6_LUNA_RESPONSES_CACHE_WRITE_12H,
GPT_5_6_RESPONSES_REASONING,
GPT_5_6_RESPONSES_STREAM,
GPT_5_6_RESPONSES_STREAM_CACHE_READ,

61
tests/integration/cost_calculation/cases/vertex_ai.py
View file

@ -3694,6 +3694,40 @@ GEMINI_3_1_PRO_PREVIEW_TIERED_CACHE_READ_ABOVE_200K: Final = _gemini_case(
},
)
GEMINI_3_1_PRO_PREVIEW_TIERED_FLEX_ABOVE_200K: Final = _gemini_case(
GEMINI_3_1_PRO_TIERED_INPUT_ABOVE_200K,
"vertex_ai/gemini-3.1-pro-preview-tiered_flex_above_200k",
"vertex_ai/gemini-3.1-pro-preview",
ExactExpected(
spend=0.456698, input_cost=0.45056, output_cost=0.006138, prompt_tokens=204800, completion_tokens=620
),
usage_fields={
"candidatesTokensDetails": [{"modality": "TEXT", "tokenCount": 620}],
"trafficType": "ON_DEMAND_FLEX",
},
)
GEMINI_3_1_PRO_PREVIEW_SERVICE_TIER_FLEX: Final = _gemini_case(
GEMINI_3_8_FLASH_SERVICE_TIER_FLEX,
"vertex_ai/gemini-3.1-pro-preview-service_tier_flex",
"vertex_ai/gemini-3.1-pro-preview",
ExactExpected(spend=0.0047432, input_cost=0.002024, output_cost=0.0027192, prompt_tokens=1840, completion_tokens=412),
usage_fields={
"candidatesTokensDetails": [{"modality": "TEXT", "tokenCount": 412}],
"trafficType": "ON_DEMAND_FLEX",
},
)
GEMINI_3_1_PRO_PREVIEW_SERVICE_TIER_PRIORITY: Final = _gemini_case(
GEMINI_3_8_FLASH_SERVICE_TIER_PRIORITY,
"vertex_ai/gemini-3.1-pro-preview-service_tier_priority",
"vertex_ai/gemini-3.1-pro-preview",
ExactExpected(
spend=0.01707552, input_cost=0.0072864, output_cost=0.00978912, prompt_tokens=1840, completion_tokens=412
),
usage_fields={
"candidatesTokensDetails": [{"modality": "TEXT", "tokenCount": 412}],
"trafficType": "ON_DEMAND_PRIORITY",
},
)
GEMINI_4_ARGON_INPUT_TEXT: Final = GEMINI_3_8_FLASH_INPUT_TEXT.model_copy(
update={
"name": "vertex_ai/gemini-4-argon-input_text",
@ -3851,6 +3885,28 @@ GEMINI_4_ARGON_TOOL_CALL: Final = GEMINI_3_8_FLASH_TOOL_CALL.model_copy(
}
)
GEMINI_4_ARGON_SERVICE_TIER_FLEX: Final = _gemini_case(
GEMINI_3_8_FLASH_SERVICE_TIER_FLEX,
"vertex_ai/gemini-4-argon-service_tier_flex",
"vertex_ai/gemini-4-argon",
ExactExpected(spend=0.00468, input_cost=0.002208, output_cost=0.002472, prompt_tokens=1840, completion_tokens=412),
usage_fields={
"candidatesTokensDetails": [{"modality": "TEXT", "tokenCount": 412}],
"trafficType": "ON_DEMAND_FLEX",
},
model_version="gemini-4-argon",
)
GEMINI_4_ARGON_SERVICE_TIER_PRIORITY: Final = _gemini_case(
GEMINI_3_8_FLASH_SERVICE_TIER_PRIORITY,
"vertex_ai/gemini-4-argon-service_tier_priority",
"vertex_ai/gemini-4-argon",
ExactExpected(spend=0.016848, input_cost=0.0079488, output_cost=0.0088992, prompt_tokens=1840, completion_tokens=412),
usage_fields={
"candidatesTokensDetails": [{"modality": "TEXT", "tokenCount": 412}],
"trafficType": "ON_DEMAND_PRIORITY",
},
model_version="gemini-4-argon",
)
GEMINI_4_ARGON_ROUTE_BRIDGES: Final = bridge_pairs(
(GEMINI_4_ARGON_INPUT_TEXT, GEMINI_4_ARGON_STREAM),
("/v1/responses", "/v1/messages"),
@ -3965,13 +4021,18 @@ CASES: Final[tuple[CostTrackingTestCase, ...]] = (
GEMINI_3_1_PRO_PREVIEW_CACHE_READ,
GEMINI_3_1_PRO_PREVIEW_TIERED_INPUT_ABOVE_200K,
GEMINI_3_1_PRO_PREVIEW_TIERED_CACHE_READ_ABOVE_200K,
GEMINI_3_1_PRO_PREVIEW_TIERED_FLEX_ABOVE_200K,
GEMINI_3_1_PRO_PREVIEW_REASONING,
GEMINI_3_1_PRO_PREVIEW_TOOL_CALL,
GEMINI_3_1_PRO_PREVIEW_SERVICE_TIER_FLEX,
GEMINI_3_1_PRO_PREVIEW_SERVICE_TIER_PRIORITY,
GEMINI_4_ARGON_INPUT_TEXT,
GEMINI_4_ARGON_STREAM,
GEMINI_4_ARGON_CACHE_READ,
GEMINI_4_ARGON_REASONING,
GEMINI_4_ARGON_TOOL_CALL,
GEMINI_4_ARGON_SERVICE_TIER_FLEX,
GEMINI_4_ARGON_SERVICE_TIER_PRIORITY,
*(pair.bridged for pair in GEMINI_4_ARGON_ROUTE_BRIDGES),
*(pair.bridged for pair in GEMINI_3_8_FLASH_ROUTE_BRIDGES),
*(pair.bridged for pair in GEMINI_3_1_PRO_PREVIEW_ROUTE_BRIDGES),

11
tests/integration/cost_calculation/cost_map.py
View file

@ -94,6 +94,12 @@ COST_MAP_ENTRIES: Final[dict[str, dict[str, JsonValue]]] = {
"cache_read_input_token_cost_above_272k_tokens_priority": 4e-08,
"cache_read_input_token_cost_above_272k_tokens": 2e-08,
"cache_read_input_token_cost_above_272k_tokens_decisions": 2e-07,
"cache_creation_input_token_cost_above_12hr": 2.5e-07,
"cache_creation_input_token_cost_above_12hr_flex": 1.25e-07,
"cache_creation_input_token_cost_above_12hr_priority": 5e-07,
"cache_creation_input_token_cost_above_12hr_above_272k_tokens": 5e-07,
"cache_creation_input_token_cost_above_12hr_above_272k_tokens_flex": 2.5e-07,
"cache_creation_input_token_cost_above_12hr_above_272k_tokens_priority": 1e-06,
"input_cost_per_token": 1e-07,
"input_cost_per_token_decisions": 1e-07,
"input_cost_per_token_above_272k_tokens": 2e-07,
@ -501,6 +507,9 @@ COST_MAP_ENTRIES: Final[dict[str, dict[str, JsonValue]]] = {
"cache_read_input_token_cost_above_200k_tokens_priority": 9.36e-07,
"cache_read_input_token_cost_flex": 2.6e-07,
"input_cost_per_token_flex": 1.3e-06,
"input_cost_per_token_above_200k_tokens_flex": 2.6e-06,
"output_cost_per_token_above_200k_tokens_flex": 1.17e-05,
"cache_read_input_token_cost_above_200k_tokens_flex": 5.2e-07,
"output_cost_per_token_flex": 7.8e-06,
},
"gemini/gemini-3.8-flash": {
@ -558,6 +567,8 @@ COST_MAP_ENTRIES: Final[dict[str, dict[str, JsonValue]]] = {
"input_cost_per_token_above_200k_tokens": 4.4e-06,
"input_cost_per_token_above_200k_tokens_priority": 7.92e-06,
"input_cost_per_token_flex": 1.1e-06,
"input_cost_per_token_above_200k_tokens_flex": 2.2e-06,
"output_cost_per_token_above_200k_tokens_flex": 9.9e-06,
"input_cost_per_token_priority": 3.96e-06,
"litellm_provider": "vertex_ai",
"max_input_tokens": 1048576,

3
tests/integration/cost_calculation/cost_tracking_case.py
View file

@ -125,6 +125,8 @@ class CostMapEntry(BaseModel):
output_cost_per_token_above_272k_tokens_ultrafast: float | None = None
input_cost_per_token_above_200k_tokens_priority: float | None = None
output_cost_per_token_above_200k_tokens_priority: float | None = None
input_cost_per_token_above_200k_tokens_flex: float | None = None
output_cost_per_token_above_200k_tokens_flex: float | None = None
cache_read_input_token_cost_flex: float | None = None
cache_read_input_token_cost_priority: float | None = None
cache_read_input_token_cost_ultrafast: float | None = None
@ -132,6 +134,7 @@ class CostMapEntry(BaseModel):
cache_read_input_token_cost_above_272k_tokens_priority: float | None = None
cache_read_input_token_cost_above_272k_tokens_ultrafast: float | None = None
cache_read_input_token_cost_above_200k_tokens_priority: float | None = None
cache_read_input_token_cost_above_200k_tokens_flex: float | None = None
cache_creation_input_token_cost_flex: float | None = None
cache_creation_input_token_cost_priority: float | None = None
cache_creation_input_token_cost_ultrafast: float | None = None