diff --git a/android/src/main/java/com/rnllama/LlamaContext.java b/android/src/main/java/com/rnllama/LlamaContext.java
index ba9b6938..414d610e 100644
--- a/android/src/main/java/com/rnllama/LlamaContext.java
+++ b/android/src/main/java/com/rnllama/LlamaContext.java
@@ -103,8 +103,8 @@ public WritableMap loadSession(String path) {
     return result;
   }
 
-  public int saveSession(String path) {
-    return saveSession(this.context, path);
+  public int saveSession(String path, int size) {
+    return saveSession(this.context, path, size);
   }
 
   public WritableMap completion(ReadableMap params) {
@@ -286,7 +286,8 @@ protected static native WritableMap loadSession(
   );
   protected static native int saveSession(
     long contextPtr,
-    String path
+    String path,
+    int size
   );
   protected static native WritableMap doCompletion(
     long context_ptr,
diff --git a/android/src/main/java/com/rnllama/RNLlama.java b/android/src/main/java/com/rnllama/RNLlama.java
index 74723db4..eb423dd0 100644
--- a/android/src/main/java/com/rnllama/RNLlama.java
+++ b/android/src/main/java/com/rnllama/RNLlama.java
@@ -112,7 +112,7 @@ protected void onPostExecute(WritableMap result) {
     tasks.put(task, "loadSession-" + contextId);
   }
 
-  public void saveSession(double id, final String path, Promise promise) {
+  public void saveSession(double id, final String path, double size, Promise promise) {
     final int contextId = (int) id;
     AsyncTask task = new AsyncTask<Void, Void, Integer>() {
       private Exception exception;
@@ -124,7 +124,7 @@ protected Integer doInBackground(Void... voids) {
           if (context == null) {
             throw new Exception("Context not found");
           }
-          Integer count = context.saveSession(path);
+          Integer count = context.saveSession(path, (int) size);
           return count;
         } catch (Exception e) {
           exception = e;
diff --git a/android/src/main/jni.cpp b/android/src/main/jni.cpp
index 7d24d9f1..5d38120a 100644
--- a/android/src/main/jni.cpp
+++ b/android/src/main/jni.cpp
@@ -222,7 +222,8 @@ Java_com_rnllama_LlamaContext_saveSession(
     JNIEnv *env,
     jobject thiz,
     jlong context_ptr,
-    jstring path
+    jstring path,
+    jint size
 ) {
     UNUSED(thiz);
     auto llama = context_map[(long) context_ptr];
@@ -230,7 +231,9 @@ Java_com_rnllama_LlamaContext_saveSession(
     const char *path_chars = env->GetStringUTFChars(path, nullptr);
 
     std::vector<llama_token> session_tokens = llama->embd;
-    if (!llama_save_session_file(llama->ctx, path_chars, session_tokens.data(), session_tokens.size())) {
+    int default_size = session_tokens.size();
+    int save_size = size > 0 && size <= default_size ? size : default_size;
+    if (!llama_save_session_file(llama->ctx, path_chars, session_tokens.data(), save_size)) {
       env->ReleaseStringUTFChars(path, path_chars);
       return -1;
     }
diff --git a/android/src/newarch/java/com/rnllama/RNLlamaModule.java b/android/src/newarch/java/com/rnllama/RNLlamaModule.java
index 55bf27e5..38c5a1c6 100644
--- a/android/src/newarch/java/com/rnllama/RNLlamaModule.java
+++ b/android/src/newarch/java/com/rnllama/RNLlamaModule.java
@@ -48,8 +48,8 @@ public void loadSession(double id, String path, Promise promise) {
   }
 
   @ReactMethod
-  public void saveSession(double id, String path, Promise promise) {
-    rnllama.saveSession(id, path, promise);
+  public void saveSession(double id, String path, double size, Promise promise) {
+    rnllama.saveSession(id, path, size, promise);
   }
 
   @ReactMethod
diff --git a/android/src/oldarch/java/com/rnllama/RNLlamaModule.java b/android/src/oldarch/java/com/rnllama/RNLlamaModule.java
index 82ee8277..4b34e2a5 100644
--- a/android/src/oldarch/java/com/rnllama/RNLlamaModule.java
+++ b/android/src/oldarch/java/com/rnllama/RNLlamaModule.java
@@ -49,8 +49,8 @@ public void loadSession(double id, String path, Promise promise) {
   }
 
   @ReactMethod
-  public void saveSession(double id, String path, Promise promise) {
-    rnllama.saveSession(id, path, promise);
+  public void saveSession(double id, String path, int size, Promise promise) {
+    rnllama.saveSession(id, path, size, promise);
   }
 
   @ReactMethod
diff --git a/ios/RNLlama.mm b/ios/RNLlama.mm
index 04089e88..8d98fe70 100644
--- a/ios/RNLlama.mm
+++ b/ios/RNLlama.mm
@@ -83,6 +83,7 @@ @implementation RNLlama
 
 RCT_EXPORT_METHOD(saveSession:(double)contextId
                  withFilePath:(NSString *)filePath
+                 withSize:(double)size
                  withResolver:(RCTPromiseResolveBlock)resolve
                  withRejecter:(RCTPromiseRejectBlock)reject)
 {
@@ -98,7 +99,7 @@ @implementation RNLlama
     dispatch_async(dispatch_get_main_queue(), ^{ // TODO: Fix for use in llamaDQue
         @try {
             @autoreleasepool {
-                int count = [context saveSession:filePath];
+                int count = [context saveSession:filePath size:(int)size];
                 resolve(@(count));
             }
         } @catch (NSException *exception) {
diff --git a/ios/RNLlamaContext.h b/ios/RNLlamaContext.h
index 93ff2eed..88a77d34 100644
--- a/ios/RNLlamaContext.h
+++ b/ios/RNLlamaContext.h
@@ -23,7 +23,7 @@
 - (NSString *)detokenize:(NSArray *)tokens;
 - (NSArray *)embedding:(NSString *)text;
 - (NSDictionary *)loadSession:(NSString *)path;
-- (int)saveSession:(NSString *)path;
+- (int)saveSession:(NSString *)path size:(int)size;
 
 - (void)invalidate;
 
diff --git a/ios/RNLlamaContext.mm b/ios/RNLlamaContext.mm
index ee5d1037..0f33570c 100644
--- a/ios/RNLlamaContext.mm
+++ b/ios/RNLlamaContext.mm
@@ -354,9 +354,11 @@ - (NSDictionary *)loadSession:(NSString *)path {
     };
 }
 
-- (int)saveSession:(NSString *)path {
+- (int)saveSession:(NSString *)path size:(int)size {
     std::vector<llama_token> session_tokens = llama->embd;
-    if (!llama_save_session_file(llama->ctx, [path UTF8String], session_tokens.data(), session_tokens.size())) {
+    int default_size = session_tokens.size();
+    int save_size = size > 0 && size <= default_size ? size : default_size;
+    if (!llama_save_session_file(llama->ctx, [path UTF8String], session_tokens.data(), save_size)) {
         @throw [NSException exceptionWithName:@"LlamaException" reason:@"Failed to save session" userInfo:nil];
     }
     return session_tokens.size();
diff --git a/src/NativeRNLlama.ts b/src/NativeRNLlama.ts
index 6554d6ee..67a949b3 100644
--- a/src/NativeRNLlama.ts
+++ b/src/NativeRNLlama.ts
@@ -116,7 +116,7 @@ export interface Spec extends TurboModule {
   initContext(params: NativeContextParams): Promise<NativeLlamaContext>;
 
   loadSession(contextId: number, filepath: string): Promise<NativeSessionLoadResult>;
-  saveSession(contextId: number, filepath: string): Promise<number>;
+  saveSession(contextId: number, filepath: string, size: number): Promise<number>;
   completion(contextId: number, params: NativeCompletionParams): Promise<NativeCompletionResult>;
   stopCompletion(contextId: number): Promise<void>;
   tokenize(contextId: number, text: string): Promise<NativeTokenizeResult>;
diff --git a/src/index.ts b/src/index.ts
index ba517d57..3150fe1c 100644
--- a/src/index.ts
+++ b/src/index.ts
@@ -67,8 +67,8 @@ export class LlamaContext {
   /**
    * Save current cached prompt & completion state to a file.
    */
-  async saveSession(filepath: string): Promise<number> {
-    return RNLlama.saveSession(this.id, filepath)
+  async saveSession(filepath: string, options?: { tokenSize: number }): Promise<number> {
+    return RNLlama.saveSession(this.id, filepath, options?.tokenSize || -1)
   }
 
   async completion(