diff --git a/koboldcpp.py b/koboldcpp.py
index 5446477a8..023f2e93e 100644
--- a/koboldcpp.py
+++ b/koboldcpp.py
@@ -644,7 +644,7 @@ if __name__ == '__main__':
     parser.add_argument("--psutil_set_threads", help="Experimental flag. If set, uses psutils to determine thread count based on physical cores.", action='store_true')
     parser.add_argument("--highpriority", help="Experimental flag. If set, increases the process CPU priority, potentially speeding up generation. Use caution.", action='store_true')
     parser.add_argument("--contextsize", help="Controls the memory allocated for maximum context size, only change if you need more RAM for big contexts. (default 2048)", type=int,choices=[512,1024,2048,4096,8192], default=2048)
-    parser.add_argument("--blasbatchsize", help="Sets the batch size used in BLAS processing (default 512)", type=int,choices=[32,64,128,256,512,1024], default=512)
+    parser.add_argument("--blasbatchsize", help="Sets the batch size used in BLAS processing (default 512, limit 256 for non-llama)", type=int,choices=[32,64,128,256,512,1024,2048], default=512)
     parser.add_argument("--stream", help="Uses pseudo streaming when generating tokens. Only for the Kobold Lite UI.", action='store_true')
     parser.add_argument("--smartcontext", help="Reserving a portion of context to try processing less frequently.", action='store_true')
     parser.add_argument("--unbantokens", help="Normally, KoboldAI prevents certain tokens such as EOS and Square Brackets. This flag unbans them.", action='store_true')