diff --git a/pkgs/development/python-modules/exllamav3/default.nix b/pkgs/development/python-modules/exllamav3/default.nix
index ab7e43ef490f..64f380cd44fe 100644
--- a/pkgs/development/python-modules/exllamav3/default.nix
+++ b/pkgs/development/python-modules/exllamav3/default.nix
@@ -7,7 +7,6 @@
 
   setuptools,
 
-  flash-attn,
   flash-linear-attention,
   formatron,
   kbnf,
@@ -22,21 +21,20 @@
   tokenizers,
   torch,
   typing-extensions,
-  xformers,
 }:
 let
   newerThanTuring = lib.filter (version: lib.versionOlder "7.9" version) torch.cudaCapabilities;
 in
 buildPythonPackage.override { inherit (torch) stdenv; } (finalAttrs: {
   pname = "exllamav3";
-  version = "0.0.43";
+  version = "1.1.0";
   pyproject = true;
 
   src = fetchFromGitHub {
     owner = "turboderp-org";
     repo = "exllamav3";
     tag = "v${finalAttrs.version}";
-    hash = "sha256-68v8ptvtOzRTnnRXrgU0emqmbCO0pECidgJ36bwm8/s=";
+    hash = "sha256-JlZt1UuTMmjaQWhiQZxzbHK3WgYYoBjP9PMEizEfsLY=";
   };
 
   pythonRelaxDeps = [
@@ -60,7 +58,6 @@ buildPythonPackage.override { inherit (torch) stdenv; } (finalAttrs: {
   ];
 
   dependencies = [
-    flash-attn
     flash-linear-attention
     formatron
     kbnf
@@ -74,7 +71,6 @@ buildPythonPackage.override { inherit (torch) stdenv; } (finalAttrs: {
     tokenizers
     torch
     typing-extensions
-    xformers
   ];
 
   env = lib.optionalAttrs torch.cudaSupport {