@meetkai/mka1 0.51.34 → 0.51.36

package/src/react-query/index.ts

@@ -139,6 +139,7 @@ export * from "./llmEvalsDeleteSchedule.js";
 export * from "./llmEvalsDeleteSuite.js";
 export * from "./llmEvalsGetArtifacts.js";
 export * from "./llmEvalsGetRun.js";
+export * from "./llmEvalsGetSampleAudio.js";
 export * from "./llmEvalsGetSchedule.js";
 export * from "./llmEvalsGetSuite.js";
 export * from "./llmEvalsGetSuiteVersion.js";
@@ -147,6 +148,7 @@ export * from "./llmEvalsListRuns.js";
 export * from "./llmEvalsListSamples.js";
 export * from "./llmEvalsListScheduleRuns.js";
 export * from "./llmEvalsListSchedules.js";
+export * from "./llmEvalsListSuiteLeaderboardRuns.js";
 export * from "./llmEvalsListSuites.js";
 export * from "./llmEvalsListSuiteVersions.js";
 export * from "./llmEvalsRerunFailedSamples.js";

package/src/react-query/llmEvalsGetSampleAudio.core.ts

@@ -0,0 +1,94 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import {
+  QueryClient,
+  QueryFunctionContext,
+  QueryKey,
+} from "@tanstack/react-query";
+import { SDKCore } from "../core.js";
+import { llmEvalsGetSampleAudio } from "../funcs/llmEvalsGetSampleAudio.js";
+import { combineSignals } from "../lib/primitives.js";
+import { RequestOptions } from "../lib/sdks.js";
+import * as components from "../models/components/index.js";
+import * as operations from "../models/operations/index.js";
+import { unwrapAsync } from "../types/fp.js";
+export type LlmEvalsGetSampleAudioQueryData = components.EvalSampleAudioObject;
+
+export function prefetchLlmEvalsGetSampleAudio(
+  queryClient: QueryClient,
+  client$: SDKCore,
+  request: operations.GetEvalSampleAudioRequest,
+  options?: RequestOptions,
+): Promise<void> {
+  return queryClient.prefetchQuery({
+    ...buildLlmEvalsGetSampleAudioQuery(
+      client$,
+      request,
+      options,
+    ),
+  });
+}
+
+export function buildLlmEvalsGetSampleAudioQuery(
+  client$: SDKCore,
+  request: operations.GetEvalSampleAudioRequest,
+  options?: RequestOptions,
+): {
+  queryKey: QueryKey;
+  queryFn: (
+    context: QueryFunctionContext,
+  ) => Promise<LlmEvalsGetSampleAudioQueryData>;
+} {
+  return {
+    queryKey: queryKeyLlmEvalsGetSampleAudio(
+      request.runId,
+      request.sampleIndex,
+      {
+        model: request.model,
+        taskId: request.taskId,
+        xOnBehalfOf: request.xOnBehalfOf,
+      },
+    ),
+    queryFn: async function llmEvalsGetSampleAudioQueryFn(
+      ctx,
+    ): Promise<LlmEvalsGetSampleAudioQueryData> {
+      const sig = combineSignals(
+        ctx.signal,
+        options?.signal,
+        options?.fetchOptions?.signal,
+      );
+      const mergedOptions = {
+        ...options?.fetchOptions,
+        ...options,
+        signal: sig,
+      };
+
+      return unwrapAsync(llmEvalsGetSampleAudio(
+        client$,
+        request,
+        mergedOptions,
+      ));
+    },
+  };
+}
+
+export function queryKeyLlmEvalsGetSampleAudio(
+  runId: string,
+  sampleIndex: number,
+  parameters: {
+    model?: string | undefined;
+    taskId?: string | undefined;
+    xOnBehalfOf?: string | undefined;
+  },
+): QueryKey {
+  return [
+    "@meetkai/mka1",
+    "evals",
+    "getSampleAudio",
+    runId,
+    sampleIndex,
+    parameters,
+  ];
+}

package/src/react-query/llmEvalsGetSampleAudio.ts

@@ -0,0 +1,154 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import {
+  InvalidateQueryFilters,
+  QueryClient,
+  useQuery,
+  UseQueryResult,
+  useSuspenseQuery,
+  UseSuspenseQueryResult,
+} from "@tanstack/react-query";
+import {
+  ConnectionError,
+  InvalidRequestError,
+  RequestAbortedError,
+  RequestTimeoutError,
+  UnexpectedClientError,
+} from "../models/errors/httpclienterrors.js";
+import { ResponseValidationError } from "../models/errors/responsevalidationerror.js";
+import { SDKError } from "../models/errors/sdkerror.js";
+import { SDKValidationError } from "../models/errors/sdkvalidationerror.js";
+import * as operations from "../models/operations/index.js";
+import { useSDKContext } from "./_context.js";
+import {
+  QueryHookOptions,
+  SuspenseQueryHookOptions,
+  TupleToPrefixes,
+} from "./_types.js";
+import {
+  buildLlmEvalsGetSampleAudioQuery,
+  LlmEvalsGetSampleAudioQueryData,
+  prefetchLlmEvalsGetSampleAudio,
+  queryKeyLlmEvalsGetSampleAudio,
+} from "./llmEvalsGetSampleAudio.core.js";
+export {
+  buildLlmEvalsGetSampleAudioQuery,
+  type LlmEvalsGetSampleAudioQueryData,
+  prefetchLlmEvalsGetSampleAudio,
+  queryKeyLlmEvalsGetSampleAudio,
+};
+
+export type LlmEvalsGetSampleAudioQueryError =
+  | SDKError
+  | ResponseValidationError
+  | ConnectionError
+  | RequestAbortedError
+  | RequestTimeoutError
+  | InvalidRequestError
+  | UnexpectedClientError
+  | SDKValidationError;
+
+/**
+ * Get eval sample audio
+ *
+ * @remarks
+ * Returns `{ object, audio, sample_index, task_id, model }` for a single transcription sample — the clip reference (a base64 `data:` URI or a URL) read from the un-redacted `dataset_row` (list responses redact inline `data:` blobs), plus the resolved task/model. Prefers the task's configured `audio_column`, falling back to a heuristic scan. The clip differs across tasks, so pass `task_id` when a run has more than one transcription task; if the index spans multiple tasks and `task_id` is omitted, returns 400 rather than guessing. `model` is optional (the clip is identical across models) and only extends the index prefix. Returns 404 when the sample or its audio is absent, 413 when the inline clip is too large.
+ */
+export function useLlmEvalsGetSampleAudio(
+  request: operations.GetEvalSampleAudioRequest,
+  options?: QueryHookOptions<
+    LlmEvalsGetSampleAudioQueryData,
+    LlmEvalsGetSampleAudioQueryError
+  >,
+): UseQueryResult<
+  LlmEvalsGetSampleAudioQueryData,
+  LlmEvalsGetSampleAudioQueryError
+> {
+  const client = useSDKContext();
+  return useQuery({
+    ...buildLlmEvalsGetSampleAudioQuery(
+      client,
+      request,
+      options,
+    ),
+    ...options,
+  });
+}
+
+/**
+ * Get eval sample audio
+ *
+ * @remarks
+ * Returns `{ object, audio, sample_index, task_id, model }` for a single transcription sample — the clip reference (a base64 `data:` URI or a URL) read from the un-redacted `dataset_row` (list responses redact inline `data:` blobs), plus the resolved task/model. Prefers the task's configured `audio_column`, falling back to a heuristic scan. The clip differs across tasks, so pass `task_id` when a run has more than one transcription task; if the index spans multiple tasks and `task_id` is omitted, returns 400 rather than guessing. `model` is optional (the clip is identical across models) and only extends the index prefix. Returns 404 when the sample or its audio is absent, 413 when the inline clip is too large.
+ */
+export function useLlmEvalsGetSampleAudioSuspense(
+  request: operations.GetEvalSampleAudioRequest,
+  options?: SuspenseQueryHookOptions<
+    LlmEvalsGetSampleAudioQueryData,
+    LlmEvalsGetSampleAudioQueryError
+  >,
+): UseSuspenseQueryResult<
+  LlmEvalsGetSampleAudioQueryData,
+  LlmEvalsGetSampleAudioQueryError
+> {
+  const client = useSDKContext();
+  return useSuspenseQuery({
+    ...buildLlmEvalsGetSampleAudioQuery(
+      client,
+      request,
+      options,
+    ),
+    ...options,
+  });
+}
+
+export function setLlmEvalsGetSampleAudioData(
+  client: QueryClient,
+  queryKeyBase: [
+    runId: string,
+    sampleIndex: number,
+    parameters: {
+      model?: string | undefined;
+      taskId?: string | undefined;
+      xOnBehalfOf?: string | undefined;
+    },
+  ],
+  data: LlmEvalsGetSampleAudioQueryData,
+): LlmEvalsGetSampleAudioQueryData | undefined {
+  const key = queryKeyLlmEvalsGetSampleAudio(...queryKeyBase);
+
+  return client.setQueryData<LlmEvalsGetSampleAudioQueryData>(key, data);
+}
+
+export function invalidateLlmEvalsGetSampleAudio(
+  client: QueryClient,
+  queryKeyBase: TupleToPrefixes<
+    [
+      runId: string,
+      sampleIndex: number,
+      parameters: {
+        model?: string | undefined;
+        taskId?: string | undefined;
+        xOnBehalfOf?: string | undefined;
+      },
+    ]
+  >,
+  filters?: Omit<InvalidateQueryFilters, "queryKey" | "predicate" | "exact">,
+): Promise<void> {
+  return client.invalidateQueries({
+    ...filters,
+    queryKey: ["@meetkai/mka1", "evals", "getSampleAudio", ...queryKeyBase],
+  });
+}
+
+export function invalidateAllLlmEvalsGetSampleAudio(
+  client: QueryClient,
+  filters?: Omit<InvalidateQueryFilters, "queryKey" | "predicate" | "exact">,
+): Promise<void> {
+  return client.invalidateQueries({
+    ...filters,
+    queryKey: ["@meetkai/mka1", "evals", "getSampleAudio"],
+  });
+}

package/src/react-query/llmEvalsListSamples.core.ts

@@ -48,6 +48,11 @@ export function buildLlmEvalsListSamplesQuery(
       taskId: request.taskId,
       model: request.model,
       status: request.status,
+      scoreMetric: request.scoreMetric,
+      scoreMin: request.scoreMin,
+      scoreMax: request.scoreMax,
+      includeDatasetRow: request.includeDatasetRow,
+      sampleIndex: request.sampleIndex,
       xOnBehalfOf: request.xOnBehalfOf,
     }),
     queryFn: async function llmEvalsListSamplesQueryFn(
@@ -81,6 +86,11 @@ export function queryKeyLlmEvalsListSamples(
     taskId?: string | undefined;
     model?: string | undefined;
     status?: operations.QueryParamStatus | undefined;
+    scoreMetric?: string | undefined;
+    scoreMin?: number | undefined;
+    scoreMax?: number | undefined;
+    includeDatasetRow?: boolean | undefined;
+    sampleIndex?: string | undefined;
     xOnBehalfOf?: string | undefined;
   },
 ): QueryKey {

package/src/react-query/llmEvalsListSamples.ts

@@ -111,6 +111,11 @@ export function setLlmEvalsListSamplesData(
       taskId?: string | undefined;
       model?: string | undefined;
       status?: operations.QueryParamStatus | undefined;
+      scoreMetric?: string | undefined;
+      scoreMin?: number | undefined;
+      scoreMax?: number | undefined;
+      includeDatasetRow?: boolean | undefined;
+      sampleIndex?: string | undefined;
       xOnBehalfOf?: string | undefined;
     },
   ],
@@ -132,6 +137,11 @@ export function invalidateLlmEvalsListSamples(
         taskId?: string | undefined;
         model?: string | undefined;
         status?: operations.QueryParamStatus | undefined;
+        scoreMetric?: string | undefined;
+        scoreMin?: number | undefined;
+        scoreMax?: number | undefined;
+        includeDatasetRow?: boolean | undefined;
+        sampleIndex?: string | undefined;
         xOnBehalfOf?: string | undefined;
       },
     ]

package/src/react-query/llmEvalsListSuiteLeaderboardRuns.core.ts

@@ -0,0 +1,92 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import {
+  QueryClient,
+  QueryFunctionContext,
+  QueryKey,
+} from "@tanstack/react-query";
+import { SDKCore } from "../core.js";
+import { llmEvalsListSuiteLeaderboardRuns } from "../funcs/llmEvalsListSuiteLeaderboardRuns.js";
+import { combineSignals } from "../lib/primitives.js";
+import { RequestOptions } from "../lib/sdks.js";
+import * as operations from "../models/operations/index.js";
+import { unwrapAsync } from "../types/fp.js";
+export type LlmEvalsListSuiteLeaderboardRunsQueryData =
+  operations.ListEvalSuiteLeaderboardRunsResponseBody;
+
+export function prefetchLlmEvalsListSuiteLeaderboardRuns(
+  queryClient: QueryClient,
+  client$: SDKCore,
+  request: operations.ListEvalSuiteLeaderboardRunsRequest,
+  options?: RequestOptions,
+): Promise<void> {
+  return queryClient.prefetchQuery({
+    ...buildLlmEvalsListSuiteLeaderboardRunsQuery(
+      client$,
+      request,
+      options,
+    ),
+  });
+}
+
+export function buildLlmEvalsListSuiteLeaderboardRunsQuery(
+  client$: SDKCore,
+  request: operations.ListEvalSuiteLeaderboardRunsRequest,
+  options?: RequestOptions,
+): {
+  queryKey: QueryKey;
+  queryFn: (
+    context: QueryFunctionContext,
+  ) => Promise<LlmEvalsListSuiteLeaderboardRunsQueryData>;
+} {
+  return {
+    queryKey: queryKeyLlmEvalsListSuiteLeaderboardRuns(request.suiteId, {
+      suiteVersion: request.suiteVersion,
+      perGroup: request.perGroup,
+      model: request.model,
+      cursor: request.cursor,
+      xOnBehalfOf: request.xOnBehalfOf,
+    }),
+    queryFn: async function llmEvalsListSuiteLeaderboardRunsQueryFn(
+      ctx,
+    ): Promise<LlmEvalsListSuiteLeaderboardRunsQueryData> {
+      const sig = combineSignals(
+        ctx.signal,
+        options?.signal,
+        options?.fetchOptions?.signal,
+      );
+      const mergedOptions = {
+        ...options?.fetchOptions,
+        ...options,
+        signal: sig,
+      };
+
+      return unwrapAsync(llmEvalsListSuiteLeaderboardRuns(
+        client$,
+        request,
+        mergedOptions,
+      ));
+    },
+  };
+}
+
+export function queryKeyLlmEvalsListSuiteLeaderboardRuns(
+  suiteId: string,
+  parameters: {
+    suiteVersion?: number | undefined;
+    perGroup?: number | undefined;
+    model?: string | undefined;
+    cursor?: string | undefined;
+    xOnBehalfOf?: string | undefined;
+  },
+): QueryKey {
+  return [
+    "@meetkai/mka1",
+    "evals",
+    "listSuiteLeaderboardRuns",
+    suiteId,
+    parameters,
+  ];
+}

package/src/react-query/llmEvalsListSuiteLeaderboardRuns.ts

@@ -0,0 +1,164 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import {
+  InvalidateQueryFilters,
+  QueryClient,
+  useQuery,
+  UseQueryResult,
+  useSuspenseQuery,
+  UseSuspenseQueryResult,
+} from "@tanstack/react-query";
+import {
+  ConnectionError,
+  InvalidRequestError,
+  RequestAbortedError,
+  RequestTimeoutError,
+  UnexpectedClientError,
+} from "../models/errors/httpclienterrors.js";
+import { ResponseValidationError } from "../models/errors/responsevalidationerror.js";
+import { SDKError } from "../models/errors/sdkerror.js";
+import { SDKValidationError } from "../models/errors/sdkvalidationerror.js";
+import * as operations from "../models/operations/index.js";
+import { useSDKContext } from "./_context.js";
+import {
+  QueryHookOptions,
+  SuspenseQueryHookOptions,
+  TupleToPrefixes,
+} from "./_types.js";
+import {
+  buildLlmEvalsListSuiteLeaderboardRunsQuery,
+  LlmEvalsListSuiteLeaderboardRunsQueryData,
+  prefetchLlmEvalsListSuiteLeaderboardRuns,
+  queryKeyLlmEvalsListSuiteLeaderboardRuns,
+} from "./llmEvalsListSuiteLeaderboardRuns.core.js";
+export {
+  buildLlmEvalsListSuiteLeaderboardRunsQuery,
+  type LlmEvalsListSuiteLeaderboardRunsQueryData,
+  prefetchLlmEvalsListSuiteLeaderboardRuns,
+  queryKeyLlmEvalsListSuiteLeaderboardRuns,
+};
+
+export type LlmEvalsListSuiteLeaderboardRunsQueryError =
+  | SDKError
+  | ResponseValidationError
+  | ConnectionError
+  | RequestAbortedError
+  | RequestTimeoutError
+  | InvalidRequestError
+  | UnexpectedClientError
+  | SDKValidationError;
+
+/**
+ * List leaderboard runs for an eval suite
+ *
+ * @remarks
+ * Returns each ranking group's most recent completed runs for a suite in one request, with the group assignments in `groups`. A group is (model, targeted task) for task_ids-restricted runs and (model) for unrestricted runs (task_id null). Build boards from `groups`, not from run.models/run.task_ids — a returned run may name models or tasks whose groups it did not rank in. `truncated: true` means a server ceiling trimmed the aggregate; whole runs are kept in (best rank, first model, first task, id) order, so every group's newest runs survive before any group's fallbacks and no group is dropped for being old. Resume a truncated board by passing `next_cursor` back as `cursor` and unioning the responses. See the evals docs for grouping rationale and cost characteristics.
+ */
+export function useLlmEvalsListSuiteLeaderboardRuns(
+  request: operations.ListEvalSuiteLeaderboardRunsRequest,
+  options?: QueryHookOptions<
+    LlmEvalsListSuiteLeaderboardRunsQueryData,
+    LlmEvalsListSuiteLeaderboardRunsQueryError
+  >,
+): UseQueryResult<
+  LlmEvalsListSuiteLeaderboardRunsQueryData,
+  LlmEvalsListSuiteLeaderboardRunsQueryError
+> {
+  const client = useSDKContext();
+  return useQuery({
+    ...buildLlmEvalsListSuiteLeaderboardRunsQuery(
+      client,
+      request,
+      options,
+    ),
+    ...options,
+  });
+}
+
+/**
+ * List leaderboard runs for an eval suite
+ *
+ * @remarks
+ * Returns each ranking group's most recent completed runs for a suite in one request, with the group assignments in `groups`. A group is (model, targeted task) for task_ids-restricted runs and (model) for unrestricted runs (task_id null). Build boards from `groups`, not from run.models/run.task_ids — a returned run may name models or tasks whose groups it did not rank in. `truncated: true` means a server ceiling trimmed the aggregate; whole runs are kept in (best rank, first model, first task, id) order, so every group's newest runs survive before any group's fallbacks and no group is dropped for being old. Resume a truncated board by passing `next_cursor` back as `cursor` and unioning the responses. See the evals docs for grouping rationale and cost characteristics.
+ */
+export function useLlmEvalsListSuiteLeaderboardRunsSuspense(
+  request: operations.ListEvalSuiteLeaderboardRunsRequest,
+  options?: SuspenseQueryHookOptions<
+    LlmEvalsListSuiteLeaderboardRunsQueryData,
+    LlmEvalsListSuiteLeaderboardRunsQueryError
+  >,
+): UseSuspenseQueryResult<
+  LlmEvalsListSuiteLeaderboardRunsQueryData,
+  LlmEvalsListSuiteLeaderboardRunsQueryError
+> {
+  const client = useSDKContext();
+  return useSuspenseQuery({
+    ...buildLlmEvalsListSuiteLeaderboardRunsQuery(
+      client,
+      request,
+      options,
+    ),
+    ...options,
+  });
+}
+
+export function setLlmEvalsListSuiteLeaderboardRunsData(
+  client: QueryClient,
+  queryKeyBase: [
+    suiteId: string,
+    parameters: {
+      suiteVersion?: number | undefined;
+      perGroup?: number | undefined;
+      model?: string | undefined;
+      cursor?: string | undefined;
+      xOnBehalfOf?: string | undefined;
+    },
+  ],
+  data: LlmEvalsListSuiteLeaderboardRunsQueryData,
+): LlmEvalsListSuiteLeaderboardRunsQueryData | undefined {
+  const key = queryKeyLlmEvalsListSuiteLeaderboardRuns(...queryKeyBase);
+
+  return client.setQueryData<LlmEvalsListSuiteLeaderboardRunsQueryData>(
+    key,
+    data,
+  );
+}
+
+export function invalidateLlmEvalsListSuiteLeaderboardRuns(
+  client: QueryClient,
+  queryKeyBase: TupleToPrefixes<
+    [
+      suiteId: string,
+      parameters: {
+        suiteVersion?: number | undefined;
+        perGroup?: number | undefined;
+        model?: string | undefined;
+        cursor?: string | undefined;
+        xOnBehalfOf?: string | undefined;
+      },
+    ]
+  >,
+  filters?: Omit<InvalidateQueryFilters, "queryKey" | "predicate" | "exact">,
+): Promise<void> {
+  return client.invalidateQueries({
+    ...filters,
+    queryKey: [
+      "@meetkai/mka1",
+      "evals",
+      "listSuiteLeaderboardRuns",
+      ...queryKeyBase,
+    ],
+  });
+}
+
+export function invalidateAllLlmEvalsListSuiteLeaderboardRuns(
+  client: QueryClient,
+  filters?: Omit<InvalidateQueryFilters, "queryKey" | "predicate" | "exact">,
+): Promise<void> {
+  return client.invalidateQueries({
+    ...filters,
+    queryKey: ["@meetkai/mka1", "evals", "listSuiteLeaderboardRuns"],
+  });
+}

package/src/sdk/evals.ts

@@ -12,6 +12,7 @@ import { llmEvalsDeleteSchedule } from "../funcs/llmEvalsDeleteSchedule.js";
 import { llmEvalsDeleteSuite } from "../funcs/llmEvalsDeleteSuite.js";
 import { llmEvalsGetArtifacts } from "../funcs/llmEvalsGetArtifacts.js";
 import { llmEvalsGetRun } from "../funcs/llmEvalsGetRun.js";
+import { llmEvalsGetSampleAudio } from "../funcs/llmEvalsGetSampleAudio.js";
 import { llmEvalsGetSchedule } from "../funcs/llmEvalsGetSchedule.js";
 import { llmEvalsGetSuite } from "../funcs/llmEvalsGetSuite.js";
 import { llmEvalsGetSuiteVersion } from "../funcs/llmEvalsGetSuiteVersion.js";
@@ -20,6 +21,7 @@ import { llmEvalsListRuns } from "../funcs/llmEvalsListRuns.js";
 import { llmEvalsListSamples } from "../funcs/llmEvalsListSamples.js";
 import { llmEvalsListScheduleRuns } from "../funcs/llmEvalsListScheduleRuns.js";
 import { llmEvalsListSchedules } from "../funcs/llmEvalsListSchedules.js";
+import { llmEvalsListSuiteLeaderboardRuns } from "../funcs/llmEvalsListSuiteLeaderboardRuns.js";
 import { llmEvalsListSuites } from "../funcs/llmEvalsListSuites.js";
 import { llmEvalsListSuiteVersions } from "../funcs/llmEvalsListSuiteVersions.js";
 import { llmEvalsRerunFailedSamples } from "../funcs/llmEvalsRerunFailedSamples.js";
@@ -151,6 +153,23 @@ export class Evals extends ClientSDK {
     ));
   }
 
+  /**
+   * List leaderboard runs for an eval suite
+   *
+   * @remarks
+   * Returns each ranking group's most recent completed runs for a suite in one request, with the group assignments in `groups`. A group is (model, targeted task) for task_ids-restricted runs and (model) for unrestricted runs (task_id null). Build boards from `groups`, not from run.models/run.task_ids — a returned run may name models or tasks whose groups it did not rank in. `truncated: true` means a server ceiling trimmed the aggregate; whole runs are kept in (best rank, first model, first task, id) order, so every group's newest runs survive before any group's fallbacks and no group is dropped for being old. Resume a truncated board by passing `next_cursor` back as `cursor` and unioning the responses. See the evals docs for grouping rationale and cost characteristics.
+   */
+  async listSuiteLeaderboardRuns(
+    request: operations.ListEvalSuiteLeaderboardRunsRequest,
+    options?: RequestOptions,
+  ): Promise<operations.ListEvalSuiteLeaderboardRunsResponseBody> {
+    return unwrapAsync(llmEvalsListSuiteLeaderboardRuns(
+      this,
+      request,
+      options,
+    ));
+  }
+
   /**
    * Create an eval run
    *
@@ -287,6 +306,23 @@ export class Evals extends ClientSDK {
     ));
   }
 
+  /**
+   * Get eval sample audio
+   *
+   * @remarks
+   * Returns `{ object, audio, sample_index, task_id, model }` for a single transcription sample — the clip reference (a base64 `data:` URI or a URL) read from the un-redacted `dataset_row` (list responses redact inline `data:` blobs), plus the resolved task/model. Prefers the task's configured `audio_column`, falling back to a heuristic scan. The clip differs across tasks, so pass `task_id` when a run has more than one transcription task; if the index spans multiple tasks and `task_id` is omitted, returns 400 rather than guessing. `model` is optional (the clip is identical across models) and only extends the index prefix. Returns 404 when the sample or its audio is absent, 413 when the inline clip is too large.
+   */
+  async getSampleAudio(
+    request: operations.GetEvalSampleAudioRequest,
+    options?: RequestOptions,
+  ): Promise<components.EvalSampleAudioObject> {
+    return unwrapAsync(llmEvalsGetSampleAudio(
+      this,
+      request,
+      options,
+    ));
+  }
+
   /**
    * Get eval run artifacts
    *