diff --git a/services/vllm/config_llama.yaml b/services/vllm/config_llama.yaml
index f1d4422..280343b 100644
--- a/services/vllm/config_llama.yaml
+++ b/services/vllm/config_llama.yaml
@@ -3,6 +3,6 @@ host: "0.0.0.0"
 port: 8001
 uvicorn-log-level: "info"
 gpu-memory-utilization: 0.95
-max-model-len: 32768
+max-model-len: 20480
 dtype: auto
 enforce-eager: false