mirror of
https://github.com/BerriAI/litellm.git
synced 2026-11-11 08:38:24 +00:00
feat: add claude mythos 5.2 to vertex ai cost map (#50940)
* Add Claude Mythos 5.2 to Vertex AI cost map Copies the vertex_ai/claude-fable-5-2 and @default rows as vertex_ai/claude-mythos-5-2 and @default. Anthropic's Mythos 5.2 model page lists Google Cloud as a platform and says it has the same specifications and pricing as Fable 5.2, and Google publishes no Mythos row yet, so the new rows cite Anthropic's page as their source. Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Bill Mythos 5.2 on Vertex through the cost path in tests Adds cost_per_token checks with cache reads, 5m and 1h cache writes and the regional uplift so the Vertex Mythos 5.2 rows are proven to charge what Fable 5.2 charges, and annotates the new test locals with Final. Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Read the regional_endpoint_uplift_multiplier key in the Vertex Mythos uplift test Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
This commit is contained in:
Kerry Lu
GitHub
parent
1035971a05
commit
0b070d25d3
3 changed files with 253 additions and 0 deletions
|
|
@ -53236,6 +53236,47 @@
|
|||
"output_cost_per_token_batches": 2e-05,
|
||||
"source": "https://cloud.google.com/gemini-enterprise-agent-platform/generative-ai/pricing"
|
||||
},
|
||||
"vertex_ai/claude-mythos-5-2": {
|
||||
"regional_endpoint_uplift_multiplier": 1.1,
|
||||
"supports_mid_conversation_system": true,
|
||||
"cache_creation_input_token_cost": 1e-05,
|
||||
"cache_creation_input_token_cost_above_1hr": 1.6e-05,
|
||||
"cache_creation_input_token_cost_batches": 5e-06,
|
||||
"cache_read_input_token_cost": 2e-07,
|
||||
"cache_read_input_token_cost_batches": 1e-07,
|
||||
"input_cost_per_token": 8e-06,
|
||||
"litellm_provider": "vertex_ai-anthropic_models",
|
||||
"max_input_tokens": 1000000,
|
||||
"max_output_tokens": 128000,
|
||||
"mode": "chat",
|
||||
"output_cost_per_token": 4e-05,
|
||||
"search_context_cost_per_query": {
|
||||
"search_context_size_high": 0.01,
|
||||
"search_context_size_low": 0.01,
|
||||
"search_context_size_medium": 0.01
|
||||
},
|
||||
"supports_adaptive_thinking": true,
|
||||
"thinking_always_on": true,
|
||||
"supports_assistant_prefill": false,
|
||||
"supports_computer_use": true,
|
||||
"supports_forced_tool_use": false,
|
||||
"supports_function_calling": true,
|
||||
"supports_native_structured_output": true,
|
||||
"supports_pdf_input": true,
|
||||
"supports_prompt_caching": true,
|
||||
"supports_reasoning": true,
|
||||
"supports_response_schema": true,
|
||||
"supports_sampling_params": false,
|
||||
"supports_tool_choice": true,
|
||||
"supports_vision": true,
|
||||
"supports_xhigh_reasoning_effort": true,
|
||||
"supports_max_reasoning_effort": true,
|
||||
"prompt_cache_min_tokens": 512,
|
||||
"deprecation_date": "2027-11-09",
|
||||
"input_cost_per_token_batches": 4e-06,
|
||||
"output_cost_per_token_batches": 2e-05,
|
||||
"source": "https://platform.claude.com/docs/en/models/mythos-5-2/overview"
|
||||
},
|
||||
"vertex_ai/claude-fable-latest": {
|
||||
"alias_of": "vertex_ai/claude-fable-5-2"
|
||||
},
|
||||
|
|
@ -53360,6 +53401,47 @@
|
|||
"output_cost_per_token_batches": 2e-05,
|
||||
"source": "https://cloud.google.com/gemini-enterprise-agent-platform/generative-ai/pricing"
|
||||
},
|
||||
"vertex_ai/claude-mythos-5-2@default": {
|
||||
"regional_endpoint_uplift_multiplier": 1.1,
|
||||
"supports_mid_conversation_system": true,
|
||||
"cache_creation_input_token_cost": 1e-05,
|
||||
"cache_creation_input_token_cost_above_1hr": 1.6e-05,
|
||||
"cache_creation_input_token_cost_batches": 5e-06,
|
||||
"cache_read_input_token_cost": 2e-07,
|
||||
"cache_read_input_token_cost_batches": 1e-07,
|
||||
"input_cost_per_token": 8e-06,
|
||||
"litellm_provider": "vertex_ai-anthropic_models",
|
||||
"max_input_tokens": 1000000,
|
||||
"max_output_tokens": 128000,
|
||||
"mode": "chat",
|
||||
"output_cost_per_token": 4e-05,
|
||||
"search_context_cost_per_query": {
|
||||
"search_context_size_high": 0.01,
|
||||
"search_context_size_low": 0.01,
|
||||
"search_context_size_medium": 0.01
|
||||
},
|
||||
"supports_adaptive_thinking": true,
|
||||
"thinking_always_on": true,
|
||||
"supports_assistant_prefill": false,
|
||||
"supports_computer_use": true,
|
||||
"supports_forced_tool_use": false,
|
||||
"supports_function_calling": true,
|
||||
"supports_native_structured_output": true,
|
||||
"supports_pdf_input": true,
|
||||
"supports_prompt_caching": true,
|
||||
"supports_reasoning": true,
|
||||
"supports_response_schema": true,
|
||||
"supports_sampling_params": false,
|
||||
"supports_tool_choice": true,
|
||||
"supports_vision": true,
|
||||
"supports_xhigh_reasoning_effort": true,
|
||||
"supports_max_reasoning_effort": true,
|
||||
"prompt_cache_min_tokens": 512,
|
||||
"deprecation_date": "2027-11-09",
|
||||
"input_cost_per_token_batches": 4e-06,
|
||||
"output_cost_per_token_batches": 2e-05,
|
||||
"source": "https://platform.claude.com/docs/en/models/mythos-5-2/overview"
|
||||
},
|
||||
"vertex_ai/claude-opus-5": {
|
||||
"deprecation_date": "2027-01-24",
|
||||
"regional_endpoint_uplift_multiplier": 1.1,
|
||||
|
|
|
|||
|
|
@ -53236,6 +53236,47 @@
|
|||
"output_cost_per_token_batches": 2e-05,
|
||||
"source": "https://cloud.google.com/gemini-enterprise-agent-platform/generative-ai/pricing"
|
||||
},
|
||||
"vertex_ai/claude-mythos-5-2": {
|
||||
"regional_endpoint_uplift_multiplier": 1.1,
|
||||
"supports_mid_conversation_system": true,
|
||||
"cache_creation_input_token_cost": 1e-05,
|
||||
"cache_creation_input_token_cost_above_1hr": 1.6e-05,
|
||||
"cache_creation_input_token_cost_batches": 5e-06,
|
||||
"cache_read_input_token_cost": 2e-07,
|
||||
"cache_read_input_token_cost_batches": 1e-07,
|
||||
"input_cost_per_token": 8e-06,
|
||||
"litellm_provider": "vertex_ai-anthropic_models",
|
||||
"max_input_tokens": 1000000,
|
||||
"max_output_tokens": 128000,
|
||||
"mode": "chat",
|
||||
"output_cost_per_token": 4e-05,
|
||||
"search_context_cost_per_query": {
|
||||
"search_context_size_high": 0.01,
|
||||
"search_context_size_low": 0.01,
|
||||
"search_context_size_medium": 0.01
|
||||
},
|
||||
"supports_adaptive_thinking": true,
|
||||
"thinking_always_on": true,
|
||||
"supports_assistant_prefill": false,
|
||||
"supports_computer_use": true,
|
||||
"supports_forced_tool_use": false,
|
||||
"supports_function_calling": true,
|
||||
"supports_native_structured_output": true,
|
||||
"supports_pdf_input": true,
|
||||
"supports_prompt_caching": true,
|
||||
"supports_reasoning": true,
|
||||
"supports_response_schema": true,
|
||||
"supports_sampling_params": false,
|
||||
"supports_tool_choice": true,
|
||||
"supports_vision": true,
|
||||
"supports_xhigh_reasoning_effort": true,
|
||||
"supports_max_reasoning_effort": true,
|
||||
"prompt_cache_min_tokens": 512,
|
||||
"deprecation_date": "2027-11-09",
|
||||
"input_cost_per_token_batches": 4e-06,
|
||||
"output_cost_per_token_batches": 2e-05,
|
||||
"source": "https://platform.claude.com/docs/en/models/mythos-5-2/overview"
|
||||
},
|
||||
"vertex_ai/claude-fable-latest": {
|
||||
"alias_of": "vertex_ai/claude-fable-5-2"
|
||||
},
|
||||
|
|
@ -53360,6 +53401,47 @@
|
|||
"output_cost_per_token_batches": 2e-05,
|
||||
"source": "https://cloud.google.com/gemini-enterprise-agent-platform/generative-ai/pricing"
|
||||
},
|
||||
"vertex_ai/claude-mythos-5-2@default": {
|
||||
"regional_endpoint_uplift_multiplier": 1.1,
|
||||
"supports_mid_conversation_system": true,
|
||||
"cache_creation_input_token_cost": 1e-05,
|
||||
"cache_creation_input_token_cost_above_1hr": 1.6e-05,
|
||||
"cache_creation_input_token_cost_batches": 5e-06,
|
||||
"cache_read_input_token_cost": 2e-07,
|
||||
"cache_read_input_token_cost_batches": 1e-07,
|
||||
"input_cost_per_token": 8e-06,
|
||||
"litellm_provider": "vertex_ai-anthropic_models",
|
||||
"max_input_tokens": 1000000,
|
||||
"max_output_tokens": 128000,
|
||||
"mode": "chat",
|
||||
"output_cost_per_token": 4e-05,
|
||||
"search_context_cost_per_query": {
|
||||
"search_context_size_high": 0.01,
|
||||
"search_context_size_low": 0.01,
|
||||
"search_context_size_medium": 0.01
|
||||
},
|
||||
"supports_adaptive_thinking": true,
|
||||
"thinking_always_on": true,
|
||||
"supports_assistant_prefill": false,
|
||||
"supports_computer_use": true,
|
||||
"supports_forced_tool_use": false,
|
||||
"supports_function_calling": true,
|
||||
"supports_native_structured_output": true,
|
||||
"supports_pdf_input": true,
|
||||
"supports_prompt_caching": true,
|
||||
"supports_reasoning": true,
|
||||
"supports_response_schema": true,
|
||||
"supports_sampling_params": false,
|
||||
"supports_tool_choice": true,
|
||||
"supports_vision": true,
|
||||
"supports_xhigh_reasoning_effort": true,
|
||||
"supports_max_reasoning_effort": true,
|
||||
"prompt_cache_min_tokens": 512,
|
||||
"deprecation_date": "2027-11-09",
|
||||
"input_cost_per_token_batches": 4e-06,
|
||||
"output_cost_per_token_batches": 2e-05,
|
||||
"source": "https://platform.claude.com/docs/en/models/mythos-5-2/overview"
|
||||
},
|
||||
"vertex_ai/claude-opus-5": {
|
||||
"deprecation_date": "2027-01-24",
|
||||
"regional_endpoint_uplift_multiplier": 1.1,
|
||||
|
|
|
|||
|
|
@ -18,6 +18,7 @@ import pytest
|
|||
import litellm
|
||||
from litellm.constants import BEDROCK_CONVERSE_MODELS
|
||||
from litellm.litellm_core_utils.get_model_cost_map import GetModelCostMap
|
||||
from litellm.types.utils import CacheCreationTokenDetails, PromptTokensDetailsWrapper, Usage
|
||||
|
||||
REPO_ROOT = os.path.join(os.path.dirname(__file__), "../..")
|
||||
|
||||
|
|
@ -172,3 +173,91 @@ def test_fable_5_2_accepts_the_speed_param_on_the_anthropic_api(local_model_cost
|
|||
|
||||
assert AnthropicModelInfo.supports_fast_mode("claude-fable-5-2", "anthropic") is True
|
||||
assert AnthropicConfig._model_supports_speed_param("claude-fable-5-2", "anthropic") is True
|
||||
|
||||
|
||||
MYTHOS_5_2_VERTEX_ROWS: Final = (
|
||||
("vertex_ai/claude-mythos-5-2", "vertex_ai/claude-fable-5-2"),
|
||||
("vertex_ai/claude-mythos-5-2@default", "vertex_ai/claude-fable-5-2@default"),
|
||||
)
|
||||
|
||||
|
||||
def _without_source(row: dict[str, object]) -> dict[str, object]:
|
||||
return {k: v for k, v in row.items() if k != "source"}
|
||||
|
||||
|
||||
def test_mythos_5_2_vertex_rows_present_in_bundled_backup() -> None:
|
||||
backup: Final = GetModelCostMap.load_local_model_cost_map()
|
||||
root: Final = _load_root_cost_map()
|
||||
for model_name, _ in MYTHOS_5_2_VERTEX_ROWS:
|
||||
assert model_name in backup, f"Missing from backup cost map: {model_name}"
|
||||
assert backup[model_name] == root[model_name], model_name
|
||||
|
||||
|
||||
@pytest.mark.parametrize(("mythos_row", "fable_row"), MYTHOS_5_2_VERTEX_ROWS)
|
||||
def test_mythos_5_2_vertex_rows_match_fable_5_2(mythos_row: str, fable_row: str) -> None:
|
||||
"""Anthropic ships Mythos 5.2 as a verified-access twin of Fable 5.2 with "the same
|
||||
specifications and pricing" (platform.claude.com/docs/en/models/mythos-5-2/overview,
|
||||
read 2026-11-10), so the Vertex rows must stay identical apart from their source."""
|
||||
root: Final = _load_root_cost_map()
|
||||
assert _without_source(root[mythos_row]) == _without_source(root[fable_row])
|
||||
assert root[mythos_row]["source"] == "https://platform.claude.com/docs/en/models/mythos-5-2/overview"
|
||||
|
||||
|
||||
@pytest.mark.parametrize("model", [row for row, _ in MYTHOS_5_2_VERTEX_ROWS])
|
||||
def test_adaptive_thinking_detected_for_mythos_5_2_on_vertex(local_model_cost_map, model: str) -> None:
|
||||
from litellm.llms.anthropic.common_utils import AnthropicModelInfo
|
||||
|
||||
assert AnthropicModelInfo.is_adaptive_thinking_model(model, "anthropic") is True
|
||||
|
||||
|
||||
def _vertex_claude_cached_usage() -> Usage:
|
||||
return Usage(
|
||||
prompt_tokens=1600,
|
||||
completion_tokens=50,
|
||||
prompt_tokens_details=PromptTokensDetailsWrapper(cached_tokens=300, cache_creation_tokens=300),
|
||||
cache_read_input_tokens=300,
|
||||
cache_creation_input_tokens=300,
|
||||
cache_creation_token_details=CacheCreationTokenDetails(
|
||||
ephemeral_5m_input_tokens=200, ephemeral_1h_input_tokens=100
|
||||
),
|
||||
)
|
||||
|
||||
|
||||
@pytest.mark.parametrize(("mythos_row", "fable_row"), MYTHOS_5_2_VERTEX_ROWS)
|
||||
def test_mythos_5_2_on_vertex_bills_like_fable_5_2(local_model_cost_map, mythos_row: str, fable_row: str) -> None:
|
||||
"""Billing through the Vertex cost path, with cache reads, 5m and 1h cache writes and the
|
||||
regional-endpoint uplift, must charge Mythos 5.2 exactly what Fable 5.2 is charged."""
|
||||
for location in ("global", "us-east5"):
|
||||
mythos_cost: Final = litellm.cost_per_token(
|
||||
model=mythos_row,
|
||||
custom_llm_provider="vertex_ai",
|
||||
usage_object=_vertex_claude_cached_usage(),
|
||||
vertex_location=location,
|
||||
)
|
||||
fable_cost: Final = litellm.cost_per_token(
|
||||
model=fable_row,
|
||||
custom_llm_provider="vertex_ai",
|
||||
usage_object=_vertex_claude_cached_usage(),
|
||||
vertex_location=location,
|
||||
)
|
||||
assert mythos_cost == fable_cost, location
|
||||
assert mythos_cost[0] > 0 and mythos_cost[1] > 0, location
|
||||
|
||||
|
||||
@pytest.mark.parametrize("model", [row for row, _ in MYTHOS_5_2_VERTEX_ROWS])
|
||||
def test_mythos_5_2_on_vertex_applies_regional_uplift(local_model_cost_map, model: str) -> None:
|
||||
uplift: Final = _load_root_cost_map()[model]["regional_endpoint_uplift_multiplier"]
|
||||
global_cost: Final = litellm.cost_per_token(
|
||||
model=model,
|
||||
custom_llm_provider="vertex_ai",
|
||||
usage_object=_vertex_claude_cached_usage(),
|
||||
vertex_location="global",
|
||||
)
|
||||
regional_cost: Final = litellm.cost_per_token(
|
||||
model=model,
|
||||
custom_llm_provider="vertex_ai",
|
||||
usage_object=_vertex_claude_cached_usage(),
|
||||
vertex_location="us-east5",
|
||||
)
|
||||
assert regional_cost[0] == pytest.approx(global_cost[0] * uplift)
|
||||
assert regional_cost[1] == pytest.approx(global_cost[1] * uplift)
|
||||
|
|
|
|||
Loading…
Add table
Reference in a new issue