diff --git a/AiPlatform/metadata/V1/Endpoint.php b/AiPlatform/metadata/V1/Endpoint.php
index cbcb49763cfa..309d31b475c1 100644
Binary files a/AiPlatform/metadata/V1/Endpoint.php and b/AiPlatform/metadata/V1/Endpoint.php differ
diff --git a/AiPlatform/src/V1/DeployedModel.php b/AiPlatform/src/V1/DeployedModel.php
index 3f33119f03f8..18123cd20d99 100644
--- a/AiPlatform/src/V1/DeployedModel.php
+++ b/AiPlatform/src/V1/DeployedModel.php
@@ -149,6 +149,12 @@ class DeployedModel extends \Google\Protobuf\Internal\Message
      * Generated from protobuf field <code>map<string, string> system_labels = 28;</code>
      */
     private $system_labels;
+    /**
+     * Optional. Spec for configuring speculative decoding.
+     *
+     * Generated from protobuf field <code>.google.cloud.aiplatform.v1.SpeculativeDecodingSpec speculative_decoding_spec = 30 [(.google.api.field_behavior) = OPTIONAL];</code>
+     */
+    protected $speculative_decoding_spec = null;
     protected $prediction_resources;
 
     /**
@@ -245,6 +251,8 @@ class DeployedModel extends \Google\Protobuf\Internal\Message
      *     @type array|\Google\Protobuf\Internal\MapField $system_labels
      *           System labels to apply to Model Garden deployments.
      *           System labels are managed by Google for internal use only.
+     *     @type \Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec $speculative_decoding_spec
+     *           Optional. Spec for configuring speculative decoding.
      * }
      */
     public function __construct($data = NULL) {
@@ -867,6 +875,42 @@ public function setSystemLabels($var)
         return $this;
     }
 
+    /**
+     * Optional. Spec for configuring speculative decoding.
+     *
+     * Generated from protobuf field <code>.google.cloud.aiplatform.v1.SpeculativeDecodingSpec speculative_decoding_spec = 30 [(.google.api.field_behavior) = OPTIONAL];</code>
+     * @return \Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec|null
+     */
+    public function getSpeculativeDecodingSpec()
+    {
+        return $this->speculative_decoding_spec;
+    }
+
+    public function hasSpeculativeDecodingSpec()
+    {
+        return isset($this->speculative_decoding_spec);
+    }
+
+    public function clearSpeculativeDecodingSpec()
+    {
+        unset($this->speculative_decoding_spec);
+    }
+
+    /**
+     * Optional. Spec for configuring speculative decoding.
+     *
+     * Generated from protobuf field <code>.google.cloud.aiplatform.v1.SpeculativeDecodingSpec speculative_decoding_spec = 30 [(.google.api.field_behavior) = OPTIONAL];</code>
+     * @param \Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec $var
+     * @return $this
+     */
+    public function setSpeculativeDecodingSpec($var)
+    {
+        GPBUtil::checkMessage($var, \Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec::class);
+        $this->speculative_decoding_spec = $var;
+
+        return $this;
+    }
+
     /**
      * @return string
      */
diff --git a/AiPlatform/src/V1/SpeculativeDecodingSpec.php b/AiPlatform/src/V1/SpeculativeDecodingSpec.php
new file mode 100644
index 000000000000..5dbb14c34336
--- /dev/null
+++ b/AiPlatform/src/V1/SpeculativeDecodingSpec.php
@@ -0,0 +1,142 @@
+<?php
+# Generated by the protocol buffer compiler.  DO NOT EDIT!
+# source: google/cloud/aiplatform/v1/endpoint.proto
+
+namespace Google\Cloud\AIPlatform\V1;
+
+use Google\Protobuf\Internal\GPBType;
+use Google\Protobuf\Internal\RepeatedField;
+use Google\Protobuf\Internal\GPBUtil;
+
+/**
+ * Configuration for Speculative Decoding.
+ *
+ * Generated from protobuf message <code>google.cloud.aiplatform.v1.SpeculativeDecodingSpec</code>
+ */
+class SpeculativeDecodingSpec extends \Google\Protobuf\Internal\Message
+{
+    /**
+     * The number of speculative tokens to generate at each step.
+     *
+     * Generated from protobuf field <code>int32 speculative_token_count = 1;</code>
+     */
+    protected $speculative_token_count = 0;
+    protected $speculation;
+
+    /**
+     * Constructor.
+     *
+     * @param array $data {
+     *     Optional. Data for populating the Message object.
+     *
+     *     @type \Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec\DraftModelSpeculation $draft_model_speculation
+     *           draft model speculation.
+     *     @type \Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec\NgramSpeculation $ngram_speculation
+     *           N-Gram speculation.
+     *     @type int $speculative_token_count
+     *           The number of speculative tokens to generate at each step.
+     * }
+     */
+    public function __construct($data = NULL) {
+        \GPBMetadata\Google\Cloud\Aiplatform\V1\Endpoint::initOnce();
+        parent::__construct($data);
+    }
+
+    /**
+     * draft model speculation.
+     *
+     * Generated from protobuf field <code>.google.cloud.aiplatform.v1.SpeculativeDecodingSpec.DraftModelSpeculation draft_model_speculation = 2;</code>
+     * @return \Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec\DraftModelSpeculation|null
+     */
+    public function getDraftModelSpeculation()
+    {
+        return $this->readOneof(2);
+    }
+
+    public function hasDraftModelSpeculation()
+    {
+        return $this->hasOneof(2);
+    }
+
+    /**
+     * draft model speculation.
+     *
+     * Generated from protobuf field <code>.google.cloud.aiplatform.v1.SpeculativeDecodingSpec.DraftModelSpeculation draft_model_speculation = 2;</code>
+     * @param \Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec\DraftModelSpeculation $var
+     * @return $this
+     */
+    public function setDraftModelSpeculation($var)
+    {
+        GPBUtil::checkMessage($var, \Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec\DraftModelSpeculation::class);
+        $this->writeOneof(2, $var);
+
+        return $this;
+    }
+
+    /**
+     * N-Gram speculation.
+     *
+     * Generated from protobuf field <code>.google.cloud.aiplatform.v1.SpeculativeDecodingSpec.NgramSpeculation ngram_speculation = 3;</code>
+     * @return \Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec\NgramSpeculation|null
+     */
+    public function getNgramSpeculation()
+    {
+        return $this->readOneof(3);
+    }
+
+    public function hasNgramSpeculation()
+    {
+        return $this->hasOneof(3);
+    }
+
+    /**
+     * N-Gram speculation.
+     *
+     * Generated from protobuf field <code>.google.cloud.aiplatform.v1.SpeculativeDecodingSpec.NgramSpeculation ngram_speculation = 3;</code>
+     * @param \Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec\NgramSpeculation $var
+     * @return $this
+     */
+    public function setNgramSpeculation($var)
+    {
+        GPBUtil::checkMessage($var, \Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec\NgramSpeculation::class);
+        $this->writeOneof(3, $var);
+
+        return $this;
+    }
+
+    /**
+     * The number of speculative tokens to generate at each step.
+     *
+     * Generated from protobuf field <code>int32 speculative_token_count = 1;</code>
+     * @return int
+     */
+    public function getSpeculativeTokenCount()
+    {
+        return $this->speculative_token_count;
+    }
+
+    /**
+     * The number of speculative tokens to generate at each step.
+     *
+     * Generated from protobuf field <code>int32 speculative_token_count = 1;</code>
+     * @param int $var
+     * @return $this
+     */
+    public function setSpeculativeTokenCount($var)
+    {
+        GPBUtil::checkInt32($var);
+        $this->speculative_token_count = $var;
+
+        return $this;
+    }
+
+    /**
+     * @return string
+     */
+    public function getSpeculation()
+    {
+        return $this->whichOneof("speculation");
+    }
+
+}
+
diff --git a/AiPlatform/src/V1/SpeculativeDecodingSpec/DraftModelSpeculation.php b/AiPlatform/src/V1/SpeculativeDecodingSpec/DraftModelSpeculation.php
new file mode 100644
index 000000000000..8096e48ea4dd
--- /dev/null
+++ b/AiPlatform/src/V1/SpeculativeDecodingSpec/DraftModelSpeculation.php
@@ -0,0 +1,69 @@
+<?php
+# Generated by the protocol buffer compiler.  DO NOT EDIT!
+# source: google/cloud/aiplatform/v1/endpoint.proto
+
+namespace Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec;
+
+use Google\Protobuf\Internal\GPBType;
+use Google\Protobuf\Internal\RepeatedField;
+use Google\Protobuf\Internal\GPBUtil;
+
+/**
+ * Draft model speculation works by using the smaller model to generate
+ * candidate tokens for speculative decoding.
+ *
+ * Generated from protobuf message <code>google.cloud.aiplatform.v1.SpeculativeDecodingSpec.DraftModelSpeculation</code>
+ */
+class DraftModelSpeculation extends \Google\Protobuf\Internal\Message
+{
+    /**
+     * Required. The resource name of the draft model.
+     *
+     * Generated from protobuf field <code>string draft_model = 1 [(.google.api.field_behavior) = REQUIRED, (.google.api.resource_reference) = {</code>
+     */
+    protected $draft_model = '';
+
+    /**
+     * Constructor.
+     *
+     * @param array $data {
+     *     Optional. Data for populating the Message object.
+     *
+     *     @type string $draft_model
+     *           Required. The resource name of the draft model.
+     * }
+     */
+    public function __construct($data = NULL) {
+        \GPBMetadata\Google\Cloud\Aiplatform\V1\Endpoint::initOnce();
+        parent::__construct($data);
+    }
+
+    /**
+     * Required. The resource name of the draft model.
+     *
+     * Generated from protobuf field <code>string draft_model = 1 [(.google.api.field_behavior) = REQUIRED, (.google.api.resource_reference) = {</code>
+     * @return string
+     */
+    public function getDraftModel()
+    {
+        return $this->draft_model;
+    }
+
+    /**
+     * Required. The resource name of the draft model.
+     *
+     * Generated from protobuf field <code>string draft_model = 1 [(.google.api.field_behavior) = REQUIRED, (.google.api.resource_reference) = {</code>
+     * @param string $var
+     * @return $this
+     */
+    public function setDraftModel($var)
+    {
+        GPBUtil::checkString($var, True);
+        $this->draft_model = $var;
+
+        return $this;
+    }
+
+}
+
+
diff --git a/AiPlatform/src/V1/SpeculativeDecodingSpec/NgramSpeculation.php b/AiPlatform/src/V1/SpeculativeDecodingSpec/NgramSpeculation.php
new file mode 100644
index 000000000000..7fd31b4bfefb
--- /dev/null
+++ b/AiPlatform/src/V1/SpeculativeDecodingSpec/NgramSpeculation.php
@@ -0,0 +1,82 @@
+<?php
+# Generated by the protocol buffer compiler.  DO NOT EDIT!
+# source: google/cloud/aiplatform/v1/endpoint.proto
+
+namespace Google\Cloud\AIPlatform\V1\SpeculativeDecodingSpec;
+
+use Google\Protobuf\Internal\GPBType;
+use Google\Protobuf\Internal\RepeatedField;
+use Google\Protobuf\Internal\GPBUtil;
+
+/**
+ * N-Gram speculation works by trying to find matching tokens in the
+ * previous prompt sequence and use those as speculation for generating
+ * new tokens.
+ *
+ * Generated from protobuf message <code>google.cloud.aiplatform.v1.SpeculativeDecodingSpec.NgramSpeculation</code>
+ */
+class NgramSpeculation extends \Google\Protobuf\Internal\Message
+{
+    /**
+     * The number of last N input tokens used as ngram to search/match
+     * against the previous prompt sequence.
+     * This is equal to the N in N-Gram.
+     * The default value is 3 if not specified.
+     *
+     * Generated from protobuf field <code>int32 ngram_size = 1;</code>
+     */
+    protected $ngram_size = 0;
+
+    /**
+     * Constructor.
+     *
+     * @param array $data {
+     *     Optional. Data for populating the Message object.
+     *
+     *     @type int $ngram_size
+     *           The number of last N input tokens used as ngram to search/match
+     *           against the previous prompt sequence.
+     *           This is equal to the N in N-Gram.
+     *           The default value is 3 if not specified.
+     * }
+     */
+    public function __construct($data = NULL) {
+        \GPBMetadata\Google\Cloud\Aiplatform\V1\Endpoint::initOnce();
+        parent::__construct($data);
+    }
+
+    /**
+     * The number of last N input tokens used as ngram to search/match
+     * against the previous prompt sequence.
+     * This is equal to the N in N-Gram.
+     * The default value is 3 if not specified.
+     *
+     * Generated from protobuf field <code>int32 ngram_size = 1;</code>
+     * @return int
+     */
+    public function getNgramSize()
+    {
+        return $this->ngram_size;
+    }
+
+    /**
+     * The number of last N input tokens used as ngram to search/match
+     * against the previous prompt sequence.
+     * This is equal to the N in N-Gram.
+     * The default value is 3 if not specified.
+     *
+     * Generated from protobuf field <code>int32 ngram_size = 1;</code>
+     * @param int $var
+     * @return $this
+     */
+    public function setNgramSize($var)
+    {
+        GPBUtil::checkInt32($var);
+        $this->ngram_size = $var;
+
+        return $this;
+    }
+
+}
+
+