scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-11-12 08:42:18 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 207
litellm/tests/integration/cost_calculation/cases/bedrock_mantle.py
devin-ai-integration[bot] cd20ecace1
test(cost): add tier 1 cost tracking cells and sync usage shapes for 2026-10-17 (#47141) 
* test(cost): add tier 1 cost tracking cells and sync usage shapes for 2026-10-17

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>

* test(cost): restore tiered cost cases and revert out of scope usage edits

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>

* test(cost): add remaining tier 1 endpoint cells for 2026-10-17

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>
2026-10-17 11:11:46 -07:00

1214 lines
48 KiB
Python
Raw Blame History

"""Bedrock Mantle cost tracking cases for openai.gpt-6-sol, a mode responses row that LiteLLM serves on /openai/v1/responses for both /v1/responses and chat clients."""
from __future__ import annotations
import json
from collections.abc import Mapping
from typing import Final
from integration.cost_calculation.cases.openai import with_usage
from integration.cost_calculation.cases.anthropic import (
CLAUDE_OPUS_5_5_CACHE_READ,
CLAUDE_OPUS_5_5_CACHE_WRITE_1H,
CLAUDE_OPUS_5_5_CACHE_WRITE_5M,
CLAUDE_OPUS_5_5_INPUT_TEXT,
CLAUDE_OPUS_5_5_MESSAGES_INPUT_TEXT,
CLAUDE_OPUS_5_5_MESSAGES_STREAM,
CLAUDE_OPUS_5_5_STREAM,
CLAUDE_OPUS_5_5_THINKING,
CLAUDE_OPUS_5_5_TOOL_CALL,
)
from integration.cost_calculation.bridge_cost_parity.case import BridgeCostParityTestCase, bridge_pairs
from integration.cost_calculation.cost_tracking_case import (
CostTrackingTestCase,
Deployment,
ExactExpected,
JsonResponse,
SseResponse,
)
from integration.cost_calculation.stream_parity.case import StreamParityTestCase, sse_frames
from pydantic import JsonValue
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON: Final = CostTrackingTestCase(
name="bedrock_mantle-openai.gpt-6-sol-responses_json",
covers="quota_management.spend_tracking.cost_matrix.logs_cost",
model="bedrock_mantle/openai.gpt-6-sol",
endpoint="/v1/responses",
request={"model": "$MODEL", "input": "cost tracking fixture", "max_output_tokens": 412, "stream": False},
response=JsonResponse(
content_type="application/json",
body={
"background": False,
"billing": {"payer": "developer"},
"completed_at": 1791861771,
"created_at": 1791861771,
"error": None,
"frequency_penalty": 0.0,
"id": "resp_$REQUEST_ID",
"incomplete_details": None,
"instructions": None,
"max_output_tokens": 64,
"max_tool_calls": None,
"metadata": {},
"model": "openai.gpt-6-sol",
"object": "response",
"output": [
{
"content": [{"annotations": [], "logprobs": [], "text": "Hello!", "type": "output_text"}],
"id": "msg_$REQUEST_ID",
"phase": "final_answer",
"role": "assistant",
"status": "completed",
"type": "message",
}
],
"parallel_tool_calls": True,
"presence_penalty": 0.0,
"previous_response_id": None,
"prompt_cache_key": None,
"prompt_cache_retention": "in_memory",
"reasoning": {"summary": None, "effort": "medium", "context": "all_turns"},
"safety_identifier": None,
"service_tier": "default",
"status": "completed",
"store": True,
"temperature": 1.0,
"text": {"format": {"type": "text"}, "verbosity": "medium"},
"tool_choice": "auto",
"tools": [],
"top_logprobs": 0,
"top_p": 0.98,
"truncation": "disabled",
"usage": {
"input_tokens": 1840,
"input_tokens_details": {"cache_write_tokens": 0, "cached_tokens": 0},
"output_tokens": 412,
"output_tokens_details": {"reasoning_tokens": 0},
"total_tokens": 2252,
},
"user": None,
"moderation": None,
},
),
expected=ExactExpected(
spend=0.00858, input_cost=0.004048, output_cost=0.004532, prompt_tokens=1840, completion_tokens=412
),
)
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_STREAM: Final = CostTrackingTestCase(
name="bedrock_mantle-openai.gpt-6-sol-responses_stream",
covers="quota_management.spend_tracking.cost_matrix.logs_cost",
model="bedrock_mantle/openai.gpt-6-sol",
endpoint="/v1/responses",
request={"model": "$MODEL", "input": "cost tracking fixture", "max_output_tokens": 412, "stream": True},
response=SseResponse(
content_type="text/event-stream",
frames=sse_frames(
{
"response": {
"background": False,
"completed_at": None,
"created_at": 1791861771,
"error": None,
"frequency_penalty": 0.0,
"id": "resp_$REQUEST_ID",
"incomplete_details": None,
"instructions": None,
"max_output_tokens": 64,
"max_tool_calls": None,
"metadata": {},
"model": "openai.gpt-6-sol",
"object": "response",
"output": [],
"parallel_tool_calls": True,
"presence_penalty": 0.0,
"previous_response_id": None,
"prompt_cache_key": None,
"prompt_cache_retention": "in_memory",
"reasoning": {"summary": None, "effort": "medium", "context": "all_turns"},
"safety_identifier": None,
"service_tier": "auto",
"status": "in_progress",
"store": True,
"temperature": 1.0,
"text": {"format": {"type": "text"}, "verbosity": "medium"},
"tool_choice": "auto",
"tools": [],
"top_logprobs": 0,
"top_p": 0.98,
"truncation": "disabled",
"usage": None,
"user": None,
"moderation": None,
},
"sequence_number": 0,
"type": "response.created",
},
{
"response": {
"background": False,
"completed_at": None,
"created_at": 1791861771,
"error": None,
"frequency_penalty": 0.0,
"id": "resp_$REQUEST_ID",
"incomplete_details": None,
"instructions": None,
"max_output_tokens": 64,
"max_tool_calls": None,
"metadata": {},
"model": "openai.gpt-6-sol",
"object": "response",
"output": [],
"parallel_tool_calls": True,
"presence_penalty": 0.0,
"previous_response_id": None,
"prompt_cache_key": None,
"prompt_cache_retention": "in_memory",
"reasoning": {"summary": None, "effort": "medium", "context": "all_turns"},
"safety_identifier": None,
"service_tier": "auto",
"status": "in_progress",
"store": True,
"temperature": 1.0,
"text": {"format": {"type": "text"}, "verbosity": "medium"},
"tool_choice": "auto",
"tools": [],
"top_logprobs": 0,
"top_p": 0.98,
"truncation": "disabled",
"usage": None,
"user": None,
"moderation": None,
},
"sequence_number": 1,
"type": "response.in_progress",
},
{
"item": {
"content": [],
"id": "msg_$REQUEST_ID",
"phase": "final_answer",
"role": "assistant",
"status": "in_progress",
"type": "message",
},
"output_index": 0,
"sequence_number": 2,
"type": "response.output_item.added",
},
{
"content_index": 0,
"item_id": "msg_$REQUEST_ID",
"output_index": 0,
"part": {"annotations": [], "logprobs": [], "text": "", "type": "output_text"},
"sequence_number": 3,
"type": "response.content_part.added",
},
{
"content_index": 0,
"delta": "Hello",
"item_id": "msg_$REQUEST_ID",
"logprobs": [],
"obfuscation": "abcdefghijklmnopqr",
"output_index": 0,
"sequence_number": 4,
"type": "response.output_text.delta",
},
{
"content_index": 0,
"delta": "!",
"item_id": "msg_$REQUEST_ID",
"logprobs": [],
"obfuscation": "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789",
"output_index": 0,
"sequence_number": 5,
"type": "response.output_text.delta",
},
{
"content_index": 0,
"item_id": "msg_$REQUEST_ID",
"logprobs": [],
"output_index": 0,
"sequence_number": 6,
"text": "Hello!",
"type": "response.output_text.done",
},
{
"content_index": 0,
"item_id": "msg_$REQUEST_ID",
"output_index": 0,
"part": {"annotations": [], "logprobs": [], "text": "Hello!", "type": "output_text"},
"sequence_number": 7,
"type": "response.content_part.done",
},
{
"item": {
"content": [{"annotations": [], "logprobs": [], "text": "Hello!", "type": "output_text"}],
"id": "msg_$REQUEST_ID",
"phase": "final_answer",
"role": "assistant",
"status": "completed",
"type": "message",
},
"output_index": 0,
"sequence_number": 8,
"type": "response.output_item.done",
},
{
"response": {
"background": False,
"completed_at": 1791861772,
"created_at": 1791861771,
"error": None,
"frequency_penalty": 0.0,
"id": "resp_$REQUEST_ID",
"incomplete_details": None,
"instructions": None,
"max_output_tokens": 64,
"max_tool_calls": None,
"metadata": {},
"model": "openai.gpt-6-sol",
"object": "response",
"output": [
{
"content": [{"annotations": [], "logprobs": [], "text": "Hello!", "type": "output_text"}],
"id": "msg_$REQUEST_ID",
"phase": "final_answer",
"role": "assistant",
"status": "completed",
"type": "message",
}
],
"parallel_tool_calls": True,
"presence_penalty": 0.0,
"previous_response_id": None,
"prompt_cache_key": None,
"prompt_cache_retention": "in_memory",
"reasoning": {"summary": None, "effort": "medium", "context": "all_turns"},
"safety_identifier": None,
"service_tier": "default",
"status": "completed",
"store": True,
"temperature": 1.0,
"text": {"format": {"type": "text"}, "verbosity": "medium"},
"tool_choice": "auto",
"tools": [],
"top_logprobs": 0,
"top_p": 0.98,
"truncation": "disabled",
"usage": {
"input_tokens": 1840,
"input_tokens_details": {"cache_write_tokens": 0, "cached_tokens": 0},
"output_tokens": 412,
"output_tokens_details": {"reasoning_tokens": 0},
"total_tokens": 2252,
},
"user": None,
"moderation": None,
},
"sequence_number": 9,
"type": "response.completed",
},
done=False,
),
),
expected=ExactExpected(
spend=0.00858, input_cost=0.004048, output_cost=0.004532, prompt_tokens=1840, completion_tokens=412
),
)
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_REASONING: Final = CostTrackingTestCase(
name="bedrock_mantle-openai.gpt-6-sol-responses_reasoning",
covers="quota_management.spend_tracking.cost_matrix.logs_cost",
model="bedrock_mantle/openai.gpt-6-sol",
endpoint="/v1/responses",
request={
"model": "$MODEL",
"input": "cost tracking fixture",
"max_output_tokens": 412,
"stream": False,
"reasoning": {"effort": "high"},
},
response=JsonResponse(
content_type="application/json",
body={
"background": False,
"billing": {"payer": "developer"},
"completed_at": 1791861773,
"created_at": 1791861772,
"error": None,
"frequency_penalty": 0.0,
"id": "resp_$REQUEST_ID",
"incomplete_details": None,
"instructions": None,
"max_output_tokens": 2048,
"max_tool_calls": None,
"metadata": {},
"model": "openai.gpt-6-sol",
"object": "response",
"output": [
{
"content": [],
"encrypted_content": "rsn_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",
"id": "rs_$REQUEST_ID",
"summary": [],
"type": "reasoning",
},
{
"content": [{"annotations": [], "logprobs": [], "text": "757", "type": "output_text"}],
"id": "msg_$REQUEST_ID",
"phase": "final_answer",
"role": "assistant",
"status": "completed",
"type": "message",
},
],
"parallel_tool_calls": True,
"presence_penalty": 0.0,
"previous_response_id": None,
"prompt_cache_key": None,
"prompt_cache_retention": "in_memory",
"reasoning": {"summary": None, "effort": "medium", "context": "all_turns"},
"safety_identifier": None,
"service_tier": "default",
"status": "completed",
"store": True,
"temperature": 1.0,
"text": {"format": {"type": "text"}, "verbosity": "medium"},
"tool_choice": "auto",
"tools": [],
"top_logprobs": 0,
"top_p": 0.98,
"truncation": "disabled",
"usage": {
"input_tokens": 1840,
"input_tokens_details": {"cache_write_tokens": 0, "cached_tokens": 0},
"output_tokens": 412,
"output_tokens_details": {"reasoning_tokens": 300},
"total_tokens": 2252,
},
"user": None,
"moderation": None,
},
),
expected=ExactExpected(
spend=0.00858, input_cost=0.004048, output_cost=0.004532, prompt_tokens=1840, completion_tokens=412
),
)
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_CHAT_JSON: Final = CostTrackingTestCase(
name="bedrock_mantle-openai.gpt-6-sol-chat_json",
covers="quota_management.spend_tracking.cost_matrix.logs_cost",
model="bedrock_mantle/openai.gpt-6-sol",
endpoint="/v1/chat/completions",
request={"model": "$MODEL", "messages": [{"role": "user", "content": "cost tracking fixture"}], "max_tokens": 412},
response=JsonResponse(
content_type="application/json",
body={
"background": False,
"billing": {"payer": "developer"},
"completed_at": 1791861772,
"created_at": 1791861772,
"error": None,
"frequency_penalty": 0.0,
"id": "resp_$REQUEST_ID",
"incomplete_details": None,
"instructions": "You are a deterministic pricing-harness assistant. Keep answers to a single short line.",
"max_output_tokens": 64,
"max_tool_calls": None,
"metadata": {},
"model": "openai.gpt-6-sol",
"object": "response",
"output": [
{
"content": [{"annotations": [], "logprobs": [], "text": "Hello.", "type": "output_text"}],
"id": "msg_$REQUEST_ID",
"phase": "final_answer",
"role": "assistant",
"status": "completed",
"type": "message",
}
],
"parallel_tool_calls": True,
"presence_penalty": 0.0,
"previous_response_id": None,
"prompt_cache_key": None,
"prompt_cache_retention": "in_memory",
"reasoning": {"summary": None, "effort": "medium", "context": "all_turns"},
"safety_identifier": None,
"service_tier": "default",
"status": "completed",
"store": True,
"temperature": 1.0,
"text": {"format": {"type": "text"}, "verbosity": "medium"},
"tool_choice": "auto",
"tools": [],
"top_logprobs": 0,
"top_p": 0.98,
"truncation": "disabled",
"usage": {
"input_tokens": 1840,
"input_tokens_details": {"cache_write_tokens": 0, "cached_tokens": 0},
"output_tokens": 412,
"output_tokens_details": {"reasoning_tokens": 0},
"total_tokens": 2252,
},
"user": None,
"moderation": None,
},
),
expected=ExactExpected(
spend=0.00858, input_cost=0.004048, output_cost=0.004532, prompt_tokens=1840, completion_tokens=412
),
)
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_CACHE_READ: Final = with_usage(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON,
"bedrock_mantle-openai.gpt-6-sol-responses_cache_read",
{
"input_tokens": 1840,
"input_tokens_details": {"cache_write_tokens": 0, "cached_tokens": 640},
"output_tokens": 412,
"output_tokens_details": {"reasoning_tokens": 0},
"total_tokens": 2252,
},
ExactExpected(
spend=0.0073128, input_cost=0.0027808, output_cost=0.004532, prompt_tokens=1840, completion_tokens=412
),
)
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_TIERED_INPUT_ABOVE_272K: Final = with_usage(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON,
"bedrock_mantle-openai.gpt-6-sol-tiered_input_above_272k",
{
"input_tokens": 300000,
"input_tokens_details": {"cache_write_tokens": 0, "cached_tokens": 0},
"output_tokens": 412,
"output_tokens_details": {"reasoning_tokens": 0},
"total_tokens": 300412,
},
ExactExpected(spend=1.326798, input_cost=1.32, output_cost=0.006798, prompt_tokens=300000, completion_tokens=412),
)
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_TIERED_CACHE_READ_ABOVE_272K: Final = with_usage(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON,
"bedrock_mantle-openai.gpt-6-sol-tiered_cache_read_above_272k",
{
"input_tokens": 300000,
"input_tokens_details": {"cache_write_tokens": 0, "cached_tokens": 200000},
"output_tokens": 412,
"output_tokens_details": {"reasoning_tokens": 0},
"total_tokens": 300412,
},
ExactExpected(spend=0.534798, input_cost=0.528, output_cost=0.006798, prompt_tokens=300000, completion_tokens=412),
)
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_TIERED_AT_272K_BOUNDARY: Final = with_usage(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON,
"bedrock_mantle-openai.gpt-6-sol-tiered_at_272k_boundary",
{
"input_tokens": 272000,
"input_tokens_details": {"cache_write_tokens": 0, "cached_tokens": 0},
"output_tokens": 412,
"output_tokens_details": {"reasoning_tokens": 0},
"total_tokens": 272412,
},
ExactExpected(spend=0.602932, input_cost=0.5984, output_cost=0.004532, prompt_tokens=272000, completion_tokens=412),
)
_MANTLE_GPT_RESPONSES_REQUEST: Final = {
"model": "$MODEL",
"input": "cost tracking fixture",
"max_output_tokens": 64,
"stream": False,
}
_MANTLE_GPT_REASONING_REQUEST: Final = {
**_MANTLE_GPT_RESPONSES_REQUEST,
"max_output_tokens": 128,
"reasoning": {"effort": "medium"},
}
_MANTLE_GPT_STREAM_REQUEST: Final = {**_MANTLE_GPT_RESPONSES_REQUEST, "stream": True}
_MANTLE_GPT_CHAT_REQUEST: Final = {
"model": "$MODEL",
"messages": [{"role": "user", "content": "cost tracking fixture"}],
"max_completion_tokens": 64,
}
_MANTLE_GPT_TOOL_REQUEST: Final = {
**_MANTLE_GPT_RESPONSES_REQUEST,
"tools": [
{
"type": "function",
"name": "get_weather",
"description": "Get the weather for a city",
"parameters": {
"type": "object",
"properties": {"city": {"type": "string"}},
"required": ["city"],
},
}
],
"tool_choice": "required",
}
_MANTLE_GPT_CACHE_WRITE_REQUEST: Final = {
**_MANTLE_GPT_RESPONSES_REQUEST,
"input": "cacheable pricing-harness prefix " * 1000,
"prompt_cache_key": "cost-cases-20261017",
"max_output_tokens": 64,
}
def _mantle_gpt_usage(
input_tokens: int,
output_tokens: int,
reasoning_tokens: int,
cached_tokens: int = 0,
cache_write_tokens: int = 0,
) -> dict[str, JsonValue]:
return {
"input_tokens": input_tokens,
"input_tokens_details": {"cache_write_tokens": cache_write_tokens, "cached_tokens": cached_tokens},
"output_tokens": output_tokens,
"output_tokens_details": {"reasoning_tokens": reasoning_tokens},
"total_tokens": input_tokens + output_tokens,
}
def _mantle_gpt_json_case(
base: CostTrackingTestCase,
provider_model: str,
shape: str,
usage: Mapping[str, JsonValue],
expected: ExactExpected,
request: Mapping[str, JsonValue],
) -> CostTrackingTestCase:
case: Final = with_usage(base, f"bedrock_mantle-{provider_model}-{shape}", usage, expected)
if not isinstance(case.response, JsonResponse):
raise TypeError(f"{case.name}: expected a JSON response fixture")
response: Final = case.response.model_copy(update={"body": {**case.response.body, "model": provider_model}})
return case.model_copy(
update={"model": f"bedrock_mantle/{provider_model}", "request": dict(request), "response": response}
)
def _mantle_gpt_stream_case(
provider_model: str, usage: Mapping[str, JsonValue], expected: ExactExpected
) -> CostTrackingTestCase:
base: Final = BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_STREAM
if not isinstance(base.response, SseResponse):
raise TypeError(f"{base.name}: expected an SSE response fixture")
old_usage: Final = json.dumps(_mantle_gpt_usage(1840, 412, 0))
new_usage: Final = json.dumps(dict(usage))
frames: Final = tuple(
frame.replace("openai.gpt-6-sol", provider_model).replace(old_usage, new_usage)
for frame in base.response.frames
)
response: Final = base.response.model_copy(update={"frames": frames})
return base.model_copy(
update={
"name": f"bedrock_mantle-{provider_model}-responses_stream",
"model": f"bedrock_mantle/{provider_model}",
"request": dict(_MANTLE_GPT_STREAM_REQUEST),
"response": response,
"expected": expected,
}
)
def _mantle_gpt_cases(
provider_model: str,
response_reasoning_tokens: int,
response_expected: ExactExpected,
reasoning_usage: Mapping[str, JsonValue],
reasoning_expected: ExactExpected,
cache_reasoning_tokens: int,
cache_expected: ExactExpected,
tiered_input_expected: ExactExpected,
tiered_cache_expected: ExactExpected,
tiered_boundary_expected: ExactExpected,
) -> tuple[CostTrackingTestCase, ...]:
response_usage: Final = _mantle_gpt_usage(9, 64, response_reasoning_tokens)
response_case: Final = _mantle_gpt_json_case(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON,
provider_model,
"responses_json",
response_usage,
response_expected,
_MANTLE_GPT_RESPONSES_REQUEST,
)
stream_case: Final = _mantle_gpt_stream_case(provider_model, response_usage, response_expected)
reasoning_case: Final = _mantle_gpt_json_case(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_REASONING,
provider_model,
"responses_reasoning",
reasoning_usage,
reasoning_expected,
_MANTLE_GPT_REASONING_REQUEST,
)
chat_case: Final = _mantle_gpt_json_case(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_CHAT_JSON,
provider_model,
"chat_json",
response_usage,
response_expected,
_MANTLE_GPT_CHAT_REQUEST,
)
chat_stream_case: Final = chat_case.model_copy(
update={
"name": f"bedrock_mantle-{provider_model}-chat_stream",
"request": {**chat_case.request, "stream": True},
"response": stream_case.response,
}
)
messages_pairs: Final = bridge_pairs((response_case, stream_case), ("/v1/messages",))
messages_json: Final = messages_pairs[0].bridged.model_copy(
update={"name": f"bedrock_mantle-{provider_model}-messages_bridged_json"}
)
messages_stream: Final = messages_pairs[1].bridged.model_copy(
update={"name": f"bedrock_mantle-{provider_model}-messages_bridged_stream"}
)
cache_case: Final = _mantle_gpt_json_case(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON,
provider_model,
"responses_cache_read",
_mantle_gpt_usage(3144, 16, cache_reasoning_tokens, cached_tokens=3142),
cache_expected,
{**_MANTLE_GPT_RESPONSES_REQUEST, "max_output_tokens": 16},
)
tiered_input: Final = _mantle_gpt_json_case(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON,
provider_model,
"tiered_input_above_272k",
_mantle_gpt_usage(300000, 412, 0),
tiered_input_expected,
{**_MANTLE_GPT_RESPONSES_REQUEST, "max_output_tokens": 412},
)
tiered_cache: Final = _mantle_gpt_json_case(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON,
provider_model,
"tiered_cache_read_above_272k",
_mantle_gpt_usage(300000, 412, 0, cached_tokens=200000),
tiered_cache_expected,
{**_MANTLE_GPT_RESPONSES_REQUEST, "max_output_tokens": 412},
)
tiered_boundary: Final = _mantle_gpt_json_case(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON,
provider_model,
"tiered_at_272k_boundary",
_mantle_gpt_usage(272000, 412, 0),
tiered_boundary_expected,
{**_MANTLE_GPT_RESPONSES_REQUEST, "max_output_tokens": 412},
)
return (
response_case,
stream_case,
reasoning_case,
chat_case,
cache_case,
tiered_input,
tiered_cache,
tiered_boundary,
chat_stream_case,
messages_json,
messages_stream,
)
def _mantle_gpt_audit3_cases(
provider_model: str,
tool_input_tokens: int,
tool_output_tokens: int,
tool_reasoning_tokens: int,
tool_expected: ExactExpected,
cache_output_tokens: int,
cache_reasoning_tokens: int,
cache_expected: ExactExpected,
) -> tuple[CostTrackingTestCase, ...]:
tool_case: Final = _mantle_gpt_json_case(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON,
provider_model,
"tool_call",
_mantle_gpt_usage(tool_input_tokens, tool_output_tokens, tool_reasoning_tokens),
tool_expected,
_MANTLE_GPT_TOOL_REQUEST,
)
if not isinstance(tool_case.response, JsonResponse):
raise TypeError(f"{tool_case.name}: expected a JSON response fixture")
tool_response: Final = tool_case.response.model_copy(
update={
"body": {
**tool_case.response.body,
"output": [
{
"arguments": '{"city":"Paris"}',
"call_id": "call_$REQUEST_ID",
"id": "fc_$REQUEST_ID",
"name": "get_weather",
"status": "completed",
"type": "function_call",
}
],
}
}
)
tool_case_with_response: Final = tool_case.model_copy(update={"response": tool_response})
cache_case: Final = _mantle_gpt_json_case(
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON,
provider_model,
"cache_write_5m",
_mantle_gpt_usage(
1667,
cache_output_tokens,
cache_reasoning_tokens,
cache_write_tokens=1665,
),
cache_expected,
_MANTLE_GPT_CACHE_WRITE_REQUEST,
)
return (tool_case_with_response, cache_case)
def _mantle_gpt_bridge(case: CostTrackingTestCase) -> BridgeCostParityTestCase:
bridged: Final = case.model_copy(
update={
"name": case.name.removesuffix("responses_json") + "chat_bridged_json",
"endpoint": "/v1/chat/completions",
"request": dict(_MANTLE_GPT_CHAT_REQUEST),
}
)
return BridgeCostParityTestCase(native=case, bridged=bridged)
_MANTLE_GPT_61_SOL_CASES: Final = _mantle_gpt_cases(
"openai.gpt-6.1-sol",
48,
ExactExpected(spend=0.00079618, input_cost=0.00002178, output_cost=0.0007744, prompt_tokens=9, completion_tokens=64),
_mantle_gpt_usage(16, 5, 0),
ExactExpected(spend=0.00009922, input_cost=0.00003872, output_cost=0.0000605, prompt_tokens=16, completion_tokens=5),
0,
ExactExpected(spend=0.000578622, input_cost=0.000385022, output_cost=0.0001936, prompt_tokens=3144, completion_tokens=16),
ExactExpected(spend=1.4594778, input_cost=1.452, output_cost=0.0074778, prompt_tokens=300000, completion_tokens=412),
ExactExpected(spend=0.5398778, input_cost=0.5324, output_cost=0.0074778, prompt_tokens=300000, completion_tokens=412),
ExactExpected(spend=0.6632252, input_cost=0.65824, output_cost=0.0049852, prompt_tokens=272000, completion_tokens=412),
)
_MANTLE_GPT_6_LUNA_CASES: Final = _mantle_gpt_cases(
"openai.gpt-6-luna",
64,
ExactExpected(spend=0.000043428, input_cost=0.000001188, output_cost=0.00004224, prompt_tokens=9, completion_tokens=64),
_mantle_gpt_usage(16, 18, 11),
ExactExpected(spend=0.000013992, input_cost=0.000002112, output_cost=0.00001188, prompt_tokens=16, completion_tokens=18),
16,
ExactExpected(spend=0.0000522984, input_cost=0.0000417384, output_cost=0.00001056, prompt_tokens=3144, completion_tokens=16),
ExactExpected(spend=0.07960788, input_cost=0.0792, output_cost=0.00040788, prompt_tokens=300000, completion_tokens=412),
ExactExpected(spend=0.03208788, input_cost=0.03168, output_cost=0.00040788, prompt_tokens=300000, completion_tokens=412),
ExactExpected(spend=0.03617592, input_cost=0.035904, output_cost=0.00027192, prompt_tokens=272000, completion_tokens=412),
)
_MANTLE_GPT_6_ASTRA_CASES: Final = _mantle_gpt_cases(
"openai.gpt-6-astra",
29,
ExactExpected(spend=0.0047047, input_cost=0.0001287, output_cost=0.004576, prompt_tokens=9, completion_tokens=64),
_mantle_gpt_usage(16, 5, 0),
ExactExpected(spend=0.0005863, input_cost=0.0002288, output_cost=0.0003575, prompt_tokens=16, completion_tokens=5),
0,
ExactExpected(spend=0.00566566, input_cost=0.00452166, output_cost=0.001144, prompt_tokens=3144, completion_tokens=16),
ExactExpected(spend=8.624187, input_cost=8.58, output_cost=0.044187, prompt_tokens=300000, completion_tokens=412),
ExactExpected(spend=3.476187, input_cost=3.432, output_cost=0.044187, prompt_tokens=300000, completion_tokens=412),
ExactExpected(spend=3.919058, input_cost=3.8896, output_cost=0.029458, prompt_tokens=272000, completion_tokens=412),
)
_MANTLE_GPT_56_TERRA_CASES: Final = _mantle_gpt_cases(
"openai.gpt-5.6-terra",
33,
ExactExpected(spend=0.00121044, input_cost=0.00002772, output_cost=0.00118272, prompt_tokens=9, completion_tokens=64),
_mantle_gpt_usage(16, 17, 10),
ExactExpected(spend=0.00036344, input_cost=0.00004928, output_cost=0.00031416, prompt_tokens=16, completion_tokens=17),
0,
ExactExpected(spend=0.001269576, input_cost=0.000973896, output_cost=0.00029568, prompt_tokens=3144, completion_tokens=16),
ExactExpected(spend=1.85942064, input_cost=1.848, output_cost=0.01142064, prompt_tokens=300000, completion_tokens=412),
ExactExpected(spend=0.75062064, input_cost=0.7392, output_cost=0.01142064, prompt_tokens=300000, completion_tokens=412),
ExactExpected(spend=0.84537376, input_cost=0.83776, output_cost=0.00761376, prompt_tokens=272000, completion_tokens=412),
)
_MANTLE_GPT_AUDIT3_CASES: Final = (
*_mantle_gpt_audit3_cases(
"openai.gpt-6.1-sol",
50,
20,
0,
ExactExpected(spend=0.000363, input_cost=0.000121, output_cost=0.000242, prompt_tokens=50, completion_tokens=20),
28,
0,
ExactExpected(
spend=0.005381765,
input_cost=0.005042965,
output_cost=0.0003388,
prompt_tokens=1667,
completion_tokens=28,
cache_creation_cost=0.005038125,
),
),
*_mantle_gpt_audit3_cases(
"openai.gpt-6-luna",
51,
18,
0,
ExactExpected(
spend=0.000018612, input_cost=0.000006732, output_cost=0.00001188, prompt_tokens=51, completion_tokens=18
),
64,
40,
ExactExpected(
spend=0.000317229,
input_cost=0.000274989,
output_cost=0.00004224,
prompt_tokens=1667,
completion_tokens=64,
cache_creation_cost=0.000274725,
),
),
*_mantle_gpt_audit3_cases(
"openai.gpt-6-astra",
51,
18,
0,
ExactExpected(
spend=0.0020163, input_cost=0.0007293, output_cost=0.001287, prompt_tokens=51, completion_tokens=18
),
27,
0,
ExactExpected(
spend=0.031720975,
input_cost=0.029790475,
output_cost=0.0019305,
prompt_tokens=1667,
completion_tokens=27,
cache_creation_cost=0.029761875,
),
),
*_mantle_gpt_audit3_cases(
"openai.gpt-5.6-terra",
51,
28,
8,
ExactExpected(
spend=0.00067452, input_cost=0.00015708, output_cost=0.00051744, prompt_tokens=51, completion_tokens=28
),
28,
0,
ExactExpected(
spend=0.00693385,
input_cost=0.00641641,
output_cost=0.00051744,
prompt_tokens=1667,
completion_tokens=28,
cache_creation_cost=0.00641025,
),
),
)
_MANTLE_CLAUDE_OPUS_MODEL: Final = "bedrock_mantle/anthropic.claude-opus-5-5"
_ADAPTIVE_USAGE: Final = {
"input_tokens": 24,
"cache_creation_input_tokens": 0,
"cache_read_input_tokens": 0,
"cache_creation": {"ephemeral_5m_input_tokens": 0, "ephemeral_1h_input_tokens": 0},
"output_tokens": 18,
"output_tokens_details": {"thinking_tokens": 6},
"service_tier": "standard",
}
_ADAPTIVE_EXPECTED: Final = ExactExpected(
spend=0.0007524, input_cost=0.0001584, output_cost=0.000594, prompt_tokens=24, completion_tokens=18
)
def _mantle_without_usage_fields(case: CostTrackingTestCase) -> CostTrackingTestCase:
if not isinstance(case.response, JsonResponse):
raise TypeError(f"{case.name}: expected a JSON response fixture")
usage: Final = case.response.body.get("usage")
if not isinstance(usage, dict):
raise TypeError(f"{case.name}: expected a usage object")
response: Final = case.response.model_copy(
update={
"body": {
**case.response.body,
"usage": {key: value for key, value in usage.items() if key not in {"inference_geo", "speed"}},
}
}
)
return case.model_copy(update={"response": response})
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_INPUT_TEXT: Final = _mantle_without_usage_fields(with_usage(
CLAUDE_OPUS_5_5_INPUT_TEXT,
"bedrock_mantle-anthropic.claude-opus-5-5-input_text",
_ADAPTIVE_USAGE,
_ADAPTIVE_EXPECTED,
).model_copy(
update={
"model": _MANTLE_CLAUDE_OPUS_MODEL,
"request": {
**CLAUDE_OPUS_5_5_INPUT_TEXT.request,
"reasoning_effort": "medium",
"allowed_openai_params": ["reasoning_effort"],
},
}
))
_MANTLE_CLAUDE_STREAM_RESPONSE: Final = SseResponse(
content_type="text/event-stream",
frames=sse_frames(
{
"type": "message_start",
"message": {
"id": "msg_$REQUEST_ID",
"type": "message",
"role": "assistant",
"model": "claude-opus-5-5",
"content": [],
"stop_reason": None,
"stop_sequence": None,
"usage": {
"input_tokens": 24,
"cache_creation_input_tokens": 0,
"cache_read_input_tokens": 0,
"cache_creation": {"ephemeral_5m_input_tokens": 0, "ephemeral_1h_input_tokens": 0},
"output_tokens": 5,
"service_tier": "standard",
},
},
},
{"type": "content_block_start", "index": 0, "content_block": {"type": "thinking", "thinking": ""}},
{"type": "content_block_delta", "index": 0, "delta": {"type": "thinking_delta", "thinking": ""}},
{"type": "content_block_stop", "index": 0},
{"type": "content_block_start", "index": 1, "content_block": {"type": "text", "text": ""}},
{"type": "content_block_delta", "index": 1, "delta": {"type": "text_delta", "text": "4"}},
{"type": "content_block_stop", "index": 1},
{
"type": "message_delta",
"delta": {"stop_reason": "end_turn", "stop_sequence": None},
"usage": {
"input_tokens": 24,
"cache_creation_input_tokens": 0,
"cache_read_input_tokens": 0,
"output_tokens": 18,
"output_tokens_details": {"thinking_tokens": 6},
},
},
{"type": "message_stop"},
),
)
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_STREAM: Final = CLAUDE_OPUS_5_5_STREAM.model_copy(
update={
"name": "bedrock_mantle-anthropic.claude-opus-5-5-stream",
"model": _MANTLE_CLAUDE_OPUS_MODEL,
"request": {
**CLAUDE_OPUS_5_5_STREAM.request,
"reasoning_effort": "medium",
"allowed_openai_params": ["reasoning_effort"],
},
"response": _MANTLE_CLAUDE_STREAM_RESPONSE,
"expected": _ADAPTIVE_EXPECTED,
}
)
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_MESSAGES_INPUT_TEXT: Final = with_usage(
CLAUDE_OPUS_5_5_MESSAGES_INPUT_TEXT,
"bedrock_mantle-anthropic.claude-opus-5-5-messages_input_text",
_ADAPTIVE_USAGE,
_ADAPTIVE_EXPECTED,
).model_copy(
update={
"model": _MANTLE_CLAUDE_OPUS_MODEL,
"request": {
**CLAUDE_OPUS_5_5_MESSAGES_INPUT_TEXT.request,
"thinking": {"type": "adaptive"},
"output_config": {"effort": "medium"},
},
}
)
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_MESSAGES_STREAM: Final = CLAUDE_OPUS_5_5_MESSAGES_STREAM.model_copy(
update={
"name": "bedrock_mantle-anthropic.claude-opus-5-5-messages_stream",
"model": _MANTLE_CLAUDE_OPUS_MODEL,
"request": {
**CLAUDE_OPUS_5_5_MESSAGES_STREAM.request,
"thinking": {"type": "adaptive"},
"output_config": {"effort": "medium"},
},
"response": _MANTLE_CLAUDE_STREAM_RESPONSE,
"expected": _ADAPTIVE_EXPECTED,
}
)
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_CACHE_READ: Final = _mantle_without_usage_fields(with_usage(
CLAUDE_OPUS_5_5_CACHE_READ,
"bedrock_mantle-anthropic.claude-opus-5-5-cache_read",
{
"input_tokens": 22,
"cache_creation_input_tokens": 0,
"cache_read_input_tokens": 4203,
"cache_creation": {"ephemeral_5m_input_tokens": 0, "ephemeral_1h_input_tokens": 0},
"output_tokens": 47,
"output_tokens_details": {"thinking_tokens": 35},
"service_tier": "standard",
},
ExactExpected(
spend=0.00308319, input_cost=0.00153219, output_cost=0.001551, prompt_tokens=4225, completion_tokens=47
),
).model_copy(update={"model": _MANTLE_CLAUDE_OPUS_MODEL}))
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_CACHE_WRITE_5M: Final = with_usage(
CLAUDE_OPUS_5_5_CACHE_WRITE_5M,
"bedrock_mantle-anthropic.claude-opus-5-5-cache_write_5m",
{
"input_tokens": 22,
"cache_creation_input_tokens": 4203,
"cache_read_input_tokens": 0,
"cache_creation": {"ephemeral_5m_input_tokens": 4203, "ephemeral_1h_input_tokens": 0},
"output_tokens": 44,
"output_tokens_details": {"thinking_tokens": 32},
"service_tier": "standard",
},
ExactExpected(
spend=0.03627195, input_cost=0.03481995, output_cost=0.001452, prompt_tokens=4225, completion_tokens=44
),
).model_copy(update={"model": _MANTLE_CLAUDE_OPUS_MODEL})
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_CACHE_WRITE_1H: Final = with_usage(
CLAUDE_OPUS_5_5_CACHE_WRITE_1H,
"bedrock_mantle-anthropic.claude-opus-5-5-cache_write_1h",
{
"input_tokens": 22,
"cache_creation_input_tokens": 9003,
"cache_read_input_tokens": 0,
"cache_creation": {"ephemeral_5m_input_tokens": 0, "ephemeral_1h_input_tokens": 9003},
"output_tokens": 12,
"output_tokens_details": {"thinking_tokens": 0},
"service_tier": "standard",
},
ExactExpected(
spend=0.1193808, input_cost=0.1189848, output_cost=0.000396, prompt_tokens=9025, completion_tokens=12
),
).model_copy(update={"model": _MANTLE_CLAUDE_OPUS_MODEL})
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_THINKING: Final = _mantle_without_usage_fields(with_usage(
CLAUDE_OPUS_5_5_THINKING,
"bedrock_mantle-anthropic.claude-opus-5-5-thinking",
_ADAPTIVE_USAGE,
_ADAPTIVE_EXPECTED,
).model_copy(update={"model": _MANTLE_CLAUDE_OPUS_MODEL}))
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_TOOL_CALL: Final = with_usage(
CLAUDE_OPUS_5_5_TOOL_CALL,
"bedrock_mantle-anthropic.claude-opus-5-5-tool_call",
{
"input_tokens": 376,
"cache_creation_input_tokens": 0,
"cache_read_input_tokens": 0,
"cache_creation": {"ephemeral_5m_input_tokens": 0, "ephemeral_1h_input_tokens": 0},
"output_tokens": 12,
"output_tokens_details": {"thinking_tokens": 0},
"service_tier": "standard",
},
ExactExpected(spend=0.0028776, input_cost=0.0024816, output_cost=0.000396, prompt_tokens=376, completion_tokens=12),
).model_copy(update={"model": _MANTLE_CLAUDE_OPUS_MODEL})
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_CHAT_BRIDGED_JSON: Final = BEDROCK_MANTLE_CLAUDE_OPUS_5_5_INPUT_TEXT.model_copy(
update={
"name": "bedrock_mantle-anthropic.claude-opus-5-5-chat_bridged_json",
"response": BEDROCK_MANTLE_CLAUDE_OPUS_5_5_MESSAGES_INPUT_TEXT.response,
}
)
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_RESPONSES_JSON: Final = BEDROCK_MANTLE_CLAUDE_OPUS_5_5_INPUT_TEXT.model_copy(
update={
"name": "bedrock_mantle-anthropic.claude-opus-5-5-responses_json",
"endpoint": "/v1/responses",
"request": {"model": "$MODEL", "input": "summarize the attached material in one line"},
}
)
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_RESPONSES_STREAM: Final = BEDROCK_MANTLE_CLAUDE_OPUS_5_5_STREAM.model_copy(
update={
"name": "bedrock_mantle-anthropic.claude-opus-5-5-responses_stream",
"endpoint": "/v1/responses",
"request": {"model": "$MODEL", "input": "summarize the attached material in one line", "stream": True},
}
)
BRIDGE_COST_PARITY: Final[tuple[BridgeCostParityTestCase, ...]] = (
_mantle_gpt_bridge(_MANTLE_GPT_61_SOL_CASES[0]),
_mantle_gpt_bridge(_MANTLE_GPT_6_LUNA_CASES[0]),
_mantle_gpt_bridge(_MANTLE_GPT_6_ASTRA_CASES[0]),
_mantle_gpt_bridge(_MANTLE_GPT_56_TERRA_CASES[0]),
BridgeCostParityTestCase(
native=BEDROCK_MANTLE_CLAUDE_OPUS_5_5_MESSAGES_INPUT_TEXT,
bridged=BEDROCK_MANTLE_CLAUDE_OPUS_5_5_CHAT_BRIDGED_JSON,
),
*tuple(
BridgeCostParityTestCase(native=cases[0], bridged=cases[9])
for cases in (_MANTLE_GPT_61_SOL_CASES, _MANTLE_GPT_6_LUNA_CASES, _MANTLE_GPT_6_ASTRA_CASES, _MANTLE_GPT_56_TERRA_CASES)
),
*tuple(
BridgeCostParityTestCase(native=cases[1], bridged=cases[10])
for cases in (_MANTLE_GPT_61_SOL_CASES, _MANTLE_GPT_6_LUNA_CASES, _MANTLE_GPT_6_ASTRA_CASES, _MANTLE_GPT_56_TERRA_CASES)
),
)
CASES: Final[tuple[CostTrackingTestCase, ...]] = (
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON,
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_STREAM,
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_REASONING,
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_CHAT_JSON,
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_CACHE_READ,
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_TIERED_INPUT_ABOVE_272K,
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_TIERED_CACHE_READ_ABOVE_272K,
BEDROCK_MANTLE_OPENAI_GPT_6_SOL_TIERED_AT_272K_BOUNDARY,
*_MANTLE_GPT_61_SOL_CASES,
*_MANTLE_GPT_6_LUNA_CASES,
*_MANTLE_GPT_6_ASTRA_CASES,
*_MANTLE_GPT_56_TERRA_CASES,
*_MANTLE_GPT_AUDIT3_CASES,
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_INPUT_TEXT,
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_STREAM,
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_MESSAGES_INPUT_TEXT,
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_MESSAGES_STREAM,
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_CHAT_BRIDGED_JSON,
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_RESPONSES_JSON,
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_RESPONSES_STREAM,
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_CACHE_READ,
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_CACHE_WRITE_5M,
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_CACHE_WRITE_1H,
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_THINKING,
BEDROCK_MANTLE_CLAUDE_OPUS_5_5_TOOL_CALL,
)
PARITY: Final[tuple[StreamParityTestCase, ...]] = (
StreamParityTestCase(
plain=BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_JSON, streamed=BEDROCK_MANTLE_OPENAI_GPT_6_SOL_RESPONSES_STREAM
),
StreamParityTestCase(plain=_MANTLE_GPT_61_SOL_CASES[0], streamed=_MANTLE_GPT_61_SOL_CASES[1]),
StreamParityTestCase(plain=_MANTLE_GPT_6_LUNA_CASES[0], streamed=_MANTLE_GPT_6_LUNA_CASES[1]),
StreamParityTestCase(plain=_MANTLE_GPT_6_ASTRA_CASES[0], streamed=_MANTLE_GPT_6_ASTRA_CASES[1]),
StreamParityTestCase(plain=_MANTLE_GPT_56_TERRA_CASES[0], streamed=_MANTLE_GPT_56_TERRA_CASES[1]),
StreamParityTestCase(
plain=BEDROCK_MANTLE_CLAUDE_OPUS_5_5_INPUT_TEXT, streamed=BEDROCK_MANTLE_CLAUDE_OPUS_5_5_STREAM
),
StreamParityTestCase(
plain=BEDROCK_MANTLE_CLAUDE_OPUS_5_5_MESSAGES_INPUT_TEXT,
streamed=BEDROCK_MANTLE_CLAUDE_OPUS_5_5_MESSAGES_STREAM,
),
StreamParityTestCase(
plain=BEDROCK_MANTLE_CLAUDE_OPUS_5_5_RESPONSES_JSON,
streamed=BEDROCK_MANTLE_CLAUDE_OPUS_5_5_RESPONSES_STREAM,
),
*tuple(
StreamParityTestCase(plain=cases[3], streamed=cases[8])
for cases in (_MANTLE_GPT_61_SOL_CASES, _MANTLE_GPT_6_LUNA_CASES, _MANTLE_GPT_6_ASTRA_CASES, _MANTLE_GPT_56_TERRA_CASES)
),
*tuple(
StreamParityTestCase(plain=cases[9], streamed=cases[10])
for cases in (_MANTLE_GPT_61_SOL_CASES, _MANTLE_GPT_6_LUNA_CASES, _MANTLE_GPT_6_ASTRA_CASES, _MANTLE_GPT_56_TERRA_CASES)
),
)
Reference in a new issue View git blame Copy permalink