diff --git a/launcher/src/main.rs b/launcher/src/main.rs
index acff85730..2e22c1007 100644
--- a/launcher/src/main.rs
+++ b/launcher/src/main.rs
@@ -158,7 +158,7 @@ fn resolve_attention(config: &Option<Config>, lora_adapters: &Option<String>) ->
                     prefix_caching = Some("0".to_string());
                 }
                 match config.model_type.as_deref() {
-                    Some("falcon") | Some("deepseek_v2") => {
+                    Some("falcon") | Some("deepseek_v2") | Some("llama4") => {
                         // Required because gemma2 needs bfloat16 which is not supported by
                         // flashinfer ?
                         if attention.is_none() {