arize-phoenix 12.2.0__py3-none-any.whl → 12.4.0__py3-none-any.whl

phoenix/server/api/routers/v1/spans.py

@@ -14,12 +14,7 @@ from pydantic import BaseModel, Field
 from sqlalchemy import exists, select, update
 from starlette.requests import Request
 from starlette.responses import Response, StreamingResponse
-from starlette.status import (
-    HTTP_202_ACCEPTED,
-    HTTP_400_BAD_REQUEST,
-    HTTP_404_NOT_FOUND,
-    HTTP_422_UNPROCESSABLE_ENTITY,
-)
+from starlette.status import HTTP_404_NOT_FOUND
 from strawberry.relay import GlobalID
 
 from phoenix.config import DEFAULT_PROJECT_NAME
@@ -33,7 +28,7 @@ from phoenix.server.api.types.node import from_global_id_with_expected_type
 from phoenix.server.authorization import is_not_locked
 from phoenix.server.bearer_auth import PhoenixUser
 from phoenix.server.dml_event import SpanAnnotationInsertEvent, SpanDeleteEvent
-from phoenix.trace.attributes import flatten
+from phoenix.trace.attributes import flatten, unflatten
 from phoenix.trace.dsl import SpanQuery as SpanQuery_
 from phoenix.trace.schemas import (
     Span as SpanForInsertion,
@@ -440,7 +435,7 @@ class SpansResponseBody(PaginatedResponseBody[Span]):
     "/spans",
     operation_id="querySpans",
     summary="Query spans with query DSL",
-    responses=add_errors_to_responses([HTTP_404_NOT_FOUND, HTTP_422_UNPROCESSABLE_ENTITY]),
+    responses=add_errors_to_responses([404, 422]),
     include_in_schema=False,
 )
 async def query_spans_handler(
@@ -467,7 +462,7 @@ async def query_spans_handler(
     except Exception as e:
         raise HTTPException(
             detail=f"Invalid query: {e}",
-            status_code=HTTP_422_UNPROCESSABLE_ENTITY,
+            status_code=422,
         )
 
     async with request.app.state.db() as session:
@@ -490,7 +485,7 @@ async def query_spans_handler(
             )
             results.append(df)
     if not results:
-        raise HTTPException(status_code=HTTP_404_NOT_FOUND)
+        raise HTTPException(status_code=404)
 
     if accept == "application/json":
         boundary_token = token_urlsafe(64)
@@ -574,7 +569,7 @@ def _to_any_value(value: Any) -> OtlpAnyValue:
     summary="Search spans with simple filters (no DSL)",
     description="Return spans within a project filtered by time range. "
     "Supports cursor-based pagination.",
-    responses=add_errors_to_responses([HTTP_404_NOT_FOUND, HTTP_422_UNPROCESSABLE_ENTITY]),
+    responses=add_errors_to_responses([404, 422]),
 )
 async def span_search_otlpv1(
     request: Request,
@@ -617,7 +612,7 @@ async def span_search_otlpv1(
             cursor_rowid = int(GlobalID.from_id(cursor).node_id)
             stmt = stmt.where(models.Span.id <= cursor_rowid)
         except Exception:
-            raise HTTPException(status_code=HTTP_422_UNPROCESSABLE_ENTITY, detail="Invalid cursor")
+            raise HTTPException(status_code=422, detail="Invalid cursor")
 
     stmt = stmt.limit(limit + 1)
 
@@ -711,7 +706,7 @@ async def span_search_otlpv1(
     summary="List spans with simple filters (no DSL)",
     description="Return spans within a project filtered by time range. "
     "Supports cursor-based pagination.",
-    responses=add_errors_to_responses([HTTP_404_NOT_FOUND, HTTP_422_UNPROCESSABLE_ENTITY]),
+    responses=add_errors_to_responses([404, 422]),
 )
 async def span_search(
     request: Request,
@@ -751,7 +746,7 @@ async def span_search(
         try:
             cursor_rowid = int(GlobalID.from_id(cursor).node_id)
         except Exception:
-            raise HTTPException(status_code=HTTP_422_UNPROCESSABLE_ENTITY, detail="Invalid cursor")
+            raise HTTPException(status_code=422, detail="Invalid cursor")
         stmt = stmt.where(models.Span.id <= cursor_rowid)
 
     stmt = stmt.limit(limit + 1)
@@ -867,9 +862,7 @@ class AnnotateSpansResponseBody(ResponseBody[list[InsertedSpanAnnotation]]):
     dependencies=[Depends(is_not_locked)],
     operation_id="annotateSpans",
     summary="Create span annotations",
-    responses=add_errors_to_responses(
-        [{"status_code": HTTP_404_NOT_FOUND, "description": "Span not found"}]
-    ),
+    responses=add_errors_to_responses([{"status_code": 404, "description": "Span not found"}]),
     response_description="Span annotations inserted successfully",
     include_in_schema=True,
 )
@@ -915,7 +908,7 @@ async def annotate_spans(
         if missing_span_ids:
             raise HTTPException(
                 detail=f"Spans with IDs {', '.join(missing_span_ids)} do not exist.",
-                status_code=HTTP_404_NOT_FOUND,
+                status_code=404,
             )
         inserted_ids = []
         dialect = SupportedSQLDialect(session.bind.dialect.name)
@@ -957,8 +950,8 @@ class CreateSpansResponseBody(V1RoutesBaseModel):
         "Submit spans to be inserted into a project. If any spans are invalid or "
         "duplicates, no spans will be inserted."
     ),
-    responses=add_errors_to_responses([HTTP_404_NOT_FOUND, HTTP_400_BAD_REQUEST]),
-    status_code=HTTP_202_ACCEPTED,
+    responses=add_errors_to_responses([404, 400]),
+    status_code=202,
 )
 async def create_spans(
     request: Request,
@@ -997,6 +990,7 @@ async def create_spans(
         # Add back the openinference.span.kind attribute since it's stored separately in the API
         attributes = dict(api_span.attributes)
         attributes["openinference.span.kind"] = api_span.span_kind
+        attributes = unflatten(attributes.items())
 
         # Create span for insertion - note we ignore the 'id' field as it's server-generated
         return SpanForInsertion(
@@ -1015,8 +1009,23 @@ async def create_spans(
             conversation=None,  # Unused
         )
 
-    async with request.app.state.db() as session:
-        project = await _get_project_by_identifier(session, project_identifier)
+    try:
+        id_ = from_global_id_with_expected_type(
+            GlobalID.from_id(project_identifier),
+            "Project",
+        )
+    except Exception:
+        project_name = project_identifier
+    else:
+        stmt = select(models.Project).filter_by(id=id_)
+        async with request.app.state.db() as session:
+            project = await session.scalar(stmt)
+        if project is None:
+            raise HTTPException(
+                status_code=HTTP_404_NOT_FOUND,
+                detail=f"Project with ID {project_identifier} not found",
+            )
+        project_name = project.name
 
     total_received = len(request_body.data)
     duplicate_spans: list[dict[str, str]] = []
@@ -1044,7 +1053,7 @@ async def create_spans(
 
         try:
             span_for_insertion = convert_api_span_for_insertion(api_span)
-            spans_to_queue.append((span_for_insertion, project.name))
+            spans_to_queue.append((span_for_insertion, project_name))
         except Exception as e:
             invalid_spans.append(
                 {
@@ -1066,7 +1075,7 @@ async def create_spans(
             "invalid_spans": invalid_spans,
         }
         raise HTTPException(
-            status_code=HTTP_400_BAD_REQUEST,
+            status_code=400,
             detail=json.dumps(error_detail),
         )
 
@@ -1102,7 +1111,7 @@ async def create_spans(
         **Note**: This operation is irreversible and may create orphaned spans.
         """
     ),
-    responses=add_errors_to_responses([HTTP_404_NOT_FOUND]),
+    responses=add_errors_to_responses([404]),
     status_code=204,  # No Content for successful deletion
 )
 async def delete_span(
@@ -1154,7 +1163,7 @@ async def delete_span(
 
         if target_span is None:
             raise HTTPException(
-                status_code=HTTP_404_NOT_FOUND,
+                status_code=404,
                 detail=error_detail,
             )
 

phoenix/server/api/routers/v1/traces.py

@@ -14,12 +14,6 @@ from starlette.concurrency import run_in_threadpool
 from starlette.datastructures import State
 from starlette.requests import Request
 from starlette.responses import Response
-from starlette.status import (
-    HTTP_404_NOT_FOUND,
-    HTTP_415_UNSUPPORTED_MEDIA_TYPE,
-    HTTP_422_UNPROCESSABLE_ENTITY,
-    HTTP_503_SERVICE_UNAVAILABLE,
-)
 from strawberry.relay import GlobalID
 
 from phoenix.db import models
@@ -49,7 +43,7 @@ def is_not_at_capacity(request: Request) -> None:
         SPAN_QUEUE_REJECTIONS.inc()
         raise HTTPException(
             detail="Server is at capacity and cannot process more requests",
-            status_code=HTTP_503_SERVICE_UNAVAILABLE,
+            status_code=503,
         )
 
 
@@ -61,14 +55,14 @@ def is_not_at_capacity(request: Request) -> None:
     responses=add_errors_to_responses(
         [
             {
-                "status_code": HTTP_415_UNSUPPORTED_MEDIA_TYPE,
+                "status_code": 415,
                 "description": (
                     "Unsupported content type (only `application/x-protobuf` is supported)"
                 ),
             },
-            {"status_code": HTTP_422_UNPROCESSABLE_ENTITY, "description": "Invalid request body"},
+            {"status_code": 422, "description": "Invalid request body"},
             {
-                "status_code": HTTP_503_SERVICE_UNAVAILABLE,
+                "status_code": 503,
                 "description": "Server is at capacity and cannot process more requests",
             },
         ]
@@ -92,12 +86,12 @@ async def post_traces(
     if content_type != "application/x-protobuf":
         raise HTTPException(
             detail=f"Unsupported content type: {content_type}",
-            status_code=HTTP_415_UNSUPPORTED_MEDIA_TYPE,
+            status_code=415,
         )
     if content_encoding and content_encoding not in ("gzip", "deflate"):
         raise HTTPException(
             detail=f"Unsupported content encoding: {content_encoding}",
-            status_code=HTTP_415_UNSUPPORTED_MEDIA_TYPE,
+            status_code=415,
         )
     body = await request.body()
     if content_encoding == "gzip":
@@ -110,7 +104,7 @@ async def post_traces(
     except DecodeError:
         raise HTTPException(
             detail="Request body is invalid ExportTraceServiceRequest",
-            status_code=HTTP_422_UNPROCESSABLE_ENTITY,
+            status_code=422,
         )
     background_tasks.add_task(_add_spans, req, request.state)
 
@@ -141,9 +135,7 @@ class AnnotateTracesResponseBody(ResponseBody[list[InsertedTraceAnnotation]]):
     dependencies=[Depends(is_not_locked)],
     operation_id="annotateTraces",
     summary="Create trace annotations",
-    responses=add_errors_to_responses(
-        [{"status_code": HTTP_404_NOT_FOUND, "description": "Trace not found"}]
-    ),
+    responses=add_errors_to_responses([{"status_code": 404, "description": "Trace not found"}]),
 )
 async def annotate_traces(
     request: Request,
@@ -177,7 +169,7 @@ async def annotate_traces(
         if missing_trace_ids:
             raise HTTPException(
                 detail=f"Traces with IDs {', '.join(missing_trace_ids)} do not exist.",
-                status_code=HTTP_404_NOT_FOUND,
+                status_code=404,
             )
         inserted_ids = []
         dialect = SupportedSQLDialect(session.bind.dialect.name)
@@ -220,7 +212,7 @@ async def _add_spans(req: ExportTraceServiceRequest, state: State) -> None:
         "2. An OpenTelemetry trace_id (hex string)\n\n"
         "This will permanently remove all spans in the trace and their associated data."
     ),
-    responses=add_errors_to_responses([HTTP_404_NOT_FOUND]),
+    responses=add_errors_to_responses([404]),
     status_code=204,  # No Content for successful deletion
 )
 async def delete_trace(
@@ -268,7 +260,7 @@ async def delete_trace(
 
         if project_id is None:
             raise HTTPException(
-                status_code=HTTP_404_NOT_FOUND,
+                status_code=404,
                 detail=error_detail,
             )
 

phoenix/server/api/routers/v1/users.py

@@ -12,15 +12,6 @@ from sqlalchemy.exc import IntegrityError as PostgreSQLIntegrityError
 from sqlalchemy.orm import joinedload
 from sqlean.dbapi2 import IntegrityError as SQLiteIntegrityError  # type: ignore[import-untyped]
 from starlette.datastructures import Secret
-from starlette.status import (
-    HTTP_201_CREATED,
-    HTTP_204_NO_CONTENT,
-    HTTP_400_BAD_REQUEST,
-    HTTP_403_FORBIDDEN,
-    HTTP_404_NOT_FOUND,
-    HTTP_409_CONFLICT,
-    HTTP_422_UNPROCESSABLE_ENTITY,
-)
 from strawberry.relay import GlobalID
 from typing_extensions import TypeAlias, assert_never
 
@@ -113,7 +104,7 @@ DEFAULT_PAGINATION_PAGE_LIMIT = 100
     response_description="A list of users.",
     responses=add_errors_to_responses(
         [
-            HTTP_422_UNPROCESSABLE_ENTITY,
+            422,
         ],
     ),
     dependencies=[Depends(require_admin)],
@@ -187,12 +178,12 @@ async def list_users(
     summary="Create a new user",
     description="Create a new user with the specified configuration.",
     response_description="The newly created user.",
-    status_code=HTTP_201_CREATED,
+    status_code=201,
     responses=add_errors_to_responses(
         [
-            {"status_code": HTTP_400_BAD_REQUEST, "description": "Role not found."},
-            {"status_code": HTTP_409_CONFLICT, "description": "Username or email already exists."},
-            HTTP_422_UNPROCESSABLE_ENTITY,
+            {"status_code": 400, "description": "Role not found."},
+            {"status_code": 409, "description": "Username or email already exists."},
+            422,
         ]
     ),
     dependencies=[Depends(require_admin), Depends(is_not_locked)],
@@ -213,14 +204,14 @@ async def create_user(
     # Prevent creation of SYSTEM users
     if role == "SYSTEM":
         raise HTTPException(
-            status_code=HTTP_400_BAD_REQUEST,
+            status_code=400,
             detail="Cannot create users with SYSTEM role",
         )
 
     # TODO: Implement VIEWER role
     if role == "VIEWER":
         raise HTTPException(
-            status_code=HTTP_400_BAD_REQUEST,
+            status_code=400,
             detail="VIEWER role not yet implemented",
         )
 
@@ -259,12 +250,12 @@ async def create_user(
             session.add(user)
     except (PostgreSQLIntegrityError, SQLiteIntegrityError) as e:
         if "users.username" in str(e):
-            raise HTTPException(status_code=HTTP_409_CONFLICT, detail="Username already exists")
+            raise HTTPException(status_code=409, detail="Username already exists")
         elif "users.email" in str(e):
-            raise HTTPException(status_code=HTTP_409_CONFLICT, detail="Email already exists")
+            raise HTTPException(status_code=409, detail="Email already exists")
         else:
             raise HTTPException(
-                status_code=HTTP_409_CONFLICT,
+                status_code=409,
                 detail="Failed to create user due to a conflict with existing data",
             )
     id_ = str(GlobalID("User", str(user.id)))
@@ -314,13 +305,13 @@ async def create_user(
     summary="Delete a user by ID",
     description="Delete an existing user by their unique GlobalID.",
     response_description="No content returned on successful deletion.",
-    status_code=HTTP_204_NO_CONTENT,
+    status_code=204,
     responses=add_errors_to_responses(
         [
-            {"status_code": HTTP_404_NOT_FOUND, "description": "User not found."},
-            HTTP_422_UNPROCESSABLE_ENTITY,
+            {"status_code": 404, "description": "User not found."},
+            422,
             {
-                "status_code": HTTP_403_FORBIDDEN,
+                "status_code": 403,
                 "description": "Cannot delete the default admin or system user",
             },
         ]

phoenix/server/api/routers/v1/utils.py

@@ -3,10 +3,6 @@ from typing import Any, Generic, Optional, TypedDict, TypeVar, Union
 from fastapi import HTTPException
 from sqlalchemy import select
 from sqlalchemy.ext.asyncio import AsyncSession
-from starlette.status import (
-    HTTP_404_NOT_FOUND,
-    HTTP_422_UNPROCESSABLE_ENTITY,
-)
 from strawberry.relay import GlobalID
 from typing_extensions import TypeAlias, assert_never
 
@@ -135,21 +131,21 @@ async def _get_project_by_identifier(
             name = project_identifier
         except HTTPException:
             raise HTTPException(
-                status_code=HTTP_422_UNPROCESSABLE_ENTITY,
+                status_code=422,
                 detail=f"Invalid project identifier format: {project_identifier}",
             )
         stmt = select(models.Project).filter_by(name=name)
         project = await session.scalar(stmt)
         if project is None:
             raise HTTPException(
-                status_code=HTTP_404_NOT_FOUND,
+                status_code=404,
                 detail=f"Project with name {name} not found",
             )
     else:
         project = await session.get(models.Project, id_)
         if project is None:
             raise HTTPException(
-                status_code=HTTP_404_NOT_FOUND,
+                status_code=404,
                 detail=f"Project with ID {project_identifier} not found",
             )
     return project

phoenix/server/app.py

@@ -45,7 +45,6 @@ from starlette.middleware.base import BaseHTTPMiddleware, RequestResponseEndpoin
 from starlette.requests import Request
 from starlette.responses import JSONResponse, PlainTextResponse, RedirectResponse, Response
 from starlette.staticfiles import StaticFiles
-from starlette.status import HTTP_401_UNAUTHORIZED
 from starlette.templating import Jinja2Templates
 from starlette.types import Scope, StatefulLifespan
 from strawberry.extensions import SchemaExtension
@@ -81,6 +80,7 @@ from phoenix.db.facilitator import Facilitator
 from phoenix.db.helpers import SupportedSQLDialect
 from phoenix.exceptions import PhoenixMigrationError
 from phoenix.pointcloud.umap_parameters import UMAPParameters
+from phoenix.server.api.auth_messages import AUTH_ERROR_MESSAGES, AuthErrorCode
 from phoenix.server.api.context import Context, DataLoaders
 from phoenix.server.api.dataloaders import (
     AnnotationConfigsByProjectDataLoader,
@@ -250,6 +250,8 @@ class AppConfig(NamedTuple):
     web_manifest_path: Path
     authentication_enabled: bool
     """ Whether authentication is enabled """
+    auth_error_messages: dict[AuthErrorCode, str]
+    """ Mapping of auth error codes to user-friendly messages """
     oauth2_idps: Sequence[OAuth2Idp]
     basic_auth_disabled: bool = False
     auto_login_idp_name: Optional[str] = None
@@ -326,6 +328,7 @@ class Static(StaticFiles):
                     "support_email": self._app_config.support_email,
                     "has_db_threshold": self._app_config.has_db_threshold,
                     "allow_external_resources": self._app_config.allow_external_resources,
+                    "auth_error_messages": self._app_config.auth_error_messages,
                 },
             )
         except Exception as e:
@@ -348,7 +351,7 @@ class RequestOriginHostnameValidator(BaseHTTPMiddleware):
             if not (url := headers.get(key)):
                 continue
             if urlparse(url).hostname not in self._trusted_hostnames:
-                return Response(f"untrusted {key}", status_code=HTTP_401_UNAUTHORIZED)
+                return Response(f"untrusted {key}", status_code=401)
         return await call_next(request)
 
 
@@ -1146,6 +1149,7 @@ def create_app(
                         and get_env_database_usage_insertion_blocking_threshold_percentage()
                     ),
                     allow_external_resources=get_env_allow_external_resources(),
+                    auth_error_messages=dict(AUTH_ERROR_MESSAGES) if authentication_enabled else {},
                 ),
             ),
             name="static",

phoenix/server/authorization.py

@@ -23,7 +23,6 @@ Usage:
 """
 
 from fastapi import HTTPException, Request
-from fastapi import status as fastapi_status
 
 from phoenix.config import get_env_support_email
 from phoenix.server.bearer_auth import PhoenixUser
@@ -49,7 +48,7 @@ def require_admin(request: Request) -> None:
     # System users have all privileges
     if not (isinstance(user, PhoenixUser) and user.is_admin):
         raise HTTPException(
-            status_code=fastapi_status.HTTP_403_FORBIDDEN,
+            status_code=403,
             detail="Only admin or system users can perform this action.",
         )
 
@@ -82,6 +81,6 @@ def is_not_locked(request: Request) -> None:
         if support_email := get_env_support_email():
             detail += f" Need help? Contact us at {support_email}"
         raise HTTPException(
-            status_code=fastapi_status.HTTP_507_INSUFFICIENT_STORAGE,
+            status_code=507,
             detail=detail,
         )

phoenix/server/bearer_auth.py

@@ -9,7 +9,6 @@ from fastapi import HTTPException, Request, WebSocket, WebSocketException
 from grpc_interceptor import AsyncServerInterceptor
 from starlette.authentication import AuthCredentials, AuthenticationBackend, BaseUser
 from starlette.requests import HTTPConnection
-from starlette.status import HTTP_401_UNAUTHORIZED
 from typing_extensions import override
 
 from phoenix import config
@@ -144,16 +143,16 @@ async def is_authenticated(
     """
     assert request or websocket
     if request and not isinstance((user := request.user), PhoenixUser):
-        raise HTTPException(status_code=HTTP_401_UNAUTHORIZED, detail="Invalid token")
+        raise HTTPException(status_code=401, detail="Invalid token")
     if websocket and not isinstance((user := websocket.user), PhoenixUser):
-        raise WebSocketException(code=HTTP_401_UNAUTHORIZED, reason="Invalid token")
+        raise WebSocketException(code=401, reason="Invalid token")
     if isinstance(user, PhoenixSystemUser):
         return
     claims = user.claims
     if claims.status is ClaimSetStatus.EXPIRED:
-        raise HTTPException(status_code=HTTP_401_UNAUTHORIZED, detail="Expired token")
+        raise HTTPException(status_code=401, detail="Expired token")
     if claims.status is not ClaimSetStatus.VALID:
-        raise HTTPException(status_code=HTTP_401_UNAUTHORIZED, detail="Invalid token")
+        raise HTTPException(status_code=401, detail="Invalid token")
 
 
 async def create_access_and_refresh_tokens(

phoenix/server/cost_tracking/model_cost_manifest.json

@@ -449,6 +449,60 @@
         }
       ]
     },
+    {
+      "name": "claude-sonnet-4-5",
+      "name_pattern": "claude-sonnet-4-5",
+      "source": "litellm",
+      "token_prices": [
+        {
+          "base_rate": 3e-6,
+          "is_prompt": true,
+          "token_type": "input"
+        },
+        {
+          "base_rate": 0.000015,
+          "is_prompt": false,
+          "token_type": "output"
+        },
+        {
+          "base_rate": 3e-7,
+          "is_prompt": true,
+          "token_type": "cache_read"
+        },
+        {
+          "base_rate": 3.75e-6,
+          "is_prompt": true,
+          "token_type": "cache_write"
+        }
+      ]
+    },
+    {
+      "name": "claude-sonnet-4-5-20250929",
+      "name_pattern": "claude-sonnet-4-5-20250929",
+      "source": "litellm",
+      "token_prices": [
+        {
+          "base_rate": 3e-6,
+          "is_prompt": true,
+          "token_type": "input"
+        },
+        {
+          "base_rate": 0.000015,
+          "is_prompt": false,
+          "token_type": "output"
+        },
+        {
+          "base_rate": 3e-7,
+          "is_prompt": true,
+          "token_type": "cache_read"
+        },
+        {
+          "base_rate": 3.75e-6,
+          "is_prompt": true,
+          "token_type": "cache_write"
+        }
+      ]
+    },
     {
       "name": "gemini-2.0-flash",
       "name_pattern": "gemini-2.0-flash(@[a-zA-Z0-9]+)?",
@@ -1028,60 +1082,6 @@
         }
       ]
     },
-    {
-      "name": "gemini-flash-latest",
-      "name_pattern": "gemini-flash-latest",
-      "source": "litellm",
-      "token_prices": [
-        {
-          "base_rate": 3e-7,
-          "is_prompt": true,
-          "token_type": "input"
-        },
-        {
-          "base_rate": 2.5e-6,
-          "is_prompt": false,
-          "token_type": "output"
-        },
-        {
-          "base_rate": 7.5e-8,
-          "is_prompt": true,
-          "token_type": "cache_read"
-        },
-        {
-          "base_rate": 1e-6,
-          "is_prompt": true,
-          "token_type": "audio"
-        }
-      ]
-    },
-    {
-      "name": "gemini-flash-lite-latest",
-      "name_pattern": "gemini-flash-lite-latest",
-      "source": "litellm",
-      "token_prices": [
-        {
-          "base_rate": 1e-7,
-          "is_prompt": true,
-          "token_type": "input"
-        },
-        {
-          "base_rate": 4e-7,
-          "is_prompt": false,
-          "token_type": "output"
-        },
-        {
-          "base_rate": 2.5e-8,
-          "is_prompt": true,
-          "token_type": "cache_read"
-        },
-        {
-          "base_rate": 3e-7,
-          "is_prompt": true,
-          "token_type": "audio"
-        }
-      ]
-    },
     {
       "name": "gpt-3.5-turbo",
       "name_pattern": "gpt-(35|3.5)-turbo",