diff --git a/docs/reference.asciidoc b/docs/reference.asciidoc
index caf4a537f..fcba0067b 100644
--- a/docs/reference.asciidoc
+++ b/docs/reference.asciidoc
@@ -6491,6 +6491,7 @@ client.ml.startTrainedModelDeployment({ model_id })
 ** *`cache_size` (Optional, number | string)*: The inference cache size (in memory outside the JVM heap) per node for the model.
 The default value is the same size as the `model_size_bytes`. To disable the cache,
 `0b` can be provided.
+** *`deployment_id` (Optional, string)*: A unique identifier for the deployment of the model.
 ** *`number_of_allocations` (Optional, number)*: The number of model allocations on each node where the model is deployed.
 All allocations on a node share the same copy of the model in memory but use
 a separate set of threads to evaluate the model.
diff --git a/src/api/types.ts b/src/api/types.ts
index 1ef86d54d..71ff546c4 100644
--- a/src/api/types.ts
+++ b/src/api/types.ts
@@ -14107,6 +14107,7 @@ export interface MlStartDatafeedResponse {
 export interface MlStartTrainedModelDeploymentRequest extends RequestBase {
   model_id: Id
   cache_size?: ByteSize
+  deployment_id?: string
   number_of_allocations?: integer
   priority?: MlTrainingPriority
   queue_capacity?: integer
diff --git a/src/api/typesWithBodyKey.ts b/src/api/typesWithBodyKey.ts
index a43ef572f..941deaea1 100644
--- a/src/api/typesWithBodyKey.ts
+++ b/src/api/typesWithBodyKey.ts
@@ -14376,6 +14376,7 @@ export interface MlStartDatafeedResponse {
 export interface MlStartTrainedModelDeploymentRequest extends RequestBase {
   model_id: Id
   cache_size?: ByteSize
+  deployment_id?: string
   number_of_allocations?: integer
   priority?: MlTrainingPriority
   queue_capacity?: integer