diff --git a/CHANGELOG.md b/CHANGELOG.md
index 769a6b7d4ad..75f0842b15a 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -39,6 +39,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Fixed `top_k` for `multiclassf1score` with one-hot encoding ([#2839](https://github.com/Lightning-AI/torchmetrics/issues/2839))
 
 
+- Fixed slow calculations of classification metrics with MPS ([#2876](https://github.com/Lightning-AI/torchmetrics/issues/2876))
+
 ---
 
 ## [1.6.0] - 2024-11-12
diff --git a/src/torchmetrics/utilities/data.py b/src/torchmetrics/utilities/data.py
index a4c80f6d4a1..dc67d5a4e34 100644
--- a/src/torchmetrics/utilities/data.py
+++ b/src/torchmetrics/utilities/data.py
@@ -199,7 +199,7 @@ def _bincount(x: Tensor, minlength: Optional[int] = None) -> Tensor:
     if minlength is None:
         minlength = len(torch.unique(x))
 
-    if torch.are_deterministic_algorithms_enabled() or _XLA_AVAILABLE and x.is_mps:
+    if torch.are_deterministic_algorithms_enabled() or _XLA_AVAILABLE or x.is_mps:
         mesh = torch.arange(minlength, device=x.device).repeat(len(x), 1)
         return torch.eq(x.reshape(-1, 1), mesh).sum(dim=0)