diff --git a/Cargo.lock b/Cargo.lock
index 8882d76..c982d96 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -215,7 +215,7 @@ dependencies = [
 
 [[package]]
 name = "axum-controller"
-version = "0.2.2"
+version = "0.2.1"
 dependencies = [
  "axum-controller-macros",
  "axum-typed-routing",
@@ -223,7 +223,7 @@ dependencies = [
 
 [[package]]
 name = "axum-controller-macros"
-version = "0.2.2"
+version = "0.2.1"
 dependencies = [
  "prettyplease",
  "proc-macro2",
diff --git a/flake.lock b/flake.lock
index 4820b73..a505047 100644
--- a/flake.lock
+++ b/flake.lock
@@ -64,16 +64,16 @@
         ]
       },
       "locked": {
-        "lastModified": 1744745160,
-        "narHash": "sha256-tTfodLDvQslNA5irDmavbOnHsCyaQ3uQKxmsoKtrKdU=",
+        "lastModified": 1745868199,
+        "narHash": "sha256-kTy1oNcN4bNHDV7yqIOSTo/JUN4enHeqyUeVE/q4iHg=",
         "owner": "ggerganov",
         "repo": "llama.cpp",
-        "rev": "80f19b41869728eeb6a26569957b92a773a2b2c6",
+        "rev": "eaea3253244dc4bbe07f6cd81325847ccc6cf93e",
         "type": "github"
       },
       "original": {
         "owner": "ggerganov",
-        "ref": "b5142",
+        "ref": "b5214",
         "repo": "llama.cpp",
         "type": "github"
       }
diff --git a/flake.nix b/flake.nix
index eac9198..e8543e5 100644
--- a/flake.nix
+++ b/flake.nix
@@ -19,7 +19,7 @@
     flake-parts.url = "github:hercules-ci/flake-parts";
     devshell.url = "github:numtide/devshell";
     llama-cpp = {
-      url = "github:ggerganov/llama.cpp/b5142";
+      url = "github:ggerganov/llama.cpp/b5214";
       inputs.nixpkgs.follows = "nixpkgs";
       inputs.flake-parts.follows = "flake-parts";
     };
diff --git a/llama_proxy_man/config.yaml b/llama_proxy_man/config.yaml
index e8048ca..9666414 100644
--- a/llama_proxy_man/config.yaml
+++ b/llama_proxy_man/config.yaml
@@ -48,12 +48,12 @@ model_specs:
     args:
       ctx-size: 16384
       flash-attn: true
-      model: /media/SanDisk/ai/models_live/mistralai_Mistral-Small-3.1-24B-Instruct-2503-IQ4_XS
+      model: /media/SanDisk/ai/models_live/Qwen_Qwen3-30B-A3B-Q5_K_S.gguf
       gpu-layers: 9999
-      model-draft: /media/SanDisk/ai/models_live/Mistral-Small-3.1-DRAFT-0.5B.Q8_0.gguf 
-      gpu-layers-draft: 9999
-      draft-max: 16
-      draft-min: 4
+      # model-draft: /media/SanDisk/ai/models_live/Qwen_Qwen3-0.6B-Q6_K.gguf
+      # gpu-layers-draft: 9999
+      # draft-max: 16
+      # draft-min: 4
   # - name: "big-chat"
   #   port: 18082
   #   vram_usage: 26.5G