diff --git a/CHANGELOG.md b/CHANGELOG.md
index 9143f002b..1a0e40fbb 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -47,6 +47,7 @@ Clipping and speaker/source alignment issues in speech separation pipeline have
 - fix(separation): fix `PixIT` training with manual optimization ([@clement-pages](https://github.com/clement-pages/))
 - fix(doc): fix link to pytorch ([@emmanuel-ferdman](https://github.com/emmanuel-ferdman/))
 - fix(task): fix corner case with small (<9) number of validation samples ([@antoinelaurent](https://github.com/antoinelaurent/))
+- fix(doc): fix default embedding in `SpeechSeparation` and `SpeakerDiarization` docstring ([@razi-tm](https://github.com/razi-tm/)).
 
 ## Version 3.3.2 (2024-09-11)
 
diff --git a/pyannote/audio/pipelines/speaker_diarization.py b/pyannote/audio/pipelines/speaker_diarization.py
index e0d43e30c..3548628e4 100644
--- a/pyannote/audio/pipelines/speaker_diarization.py
+++ b/pyannote/audio/pipelines/speaker_diarization.py
@@ -69,7 +69,7 @@ class SpeakerDiarization(SpeakerDiarizationMixin, Pipeline):
         `segmentation_step` controls the step of this window, provided as a ratio of its
         duration. Defaults to 0.1 (i.e. 90% overlap between two consecuive windows).
     embedding : Model, str, or dict, optional
-        Pretrained embedding model. Defaults to "pyannote/embedding@2022.07".
+        Pretrained embedding model. Defaults to "speechbrain/spkrec-ecapa-voxceleb@5c0be38".
         See pyannote.audio.pipelines.utils.get_model for supported format.
     embedding_exclude_overlap : bool, optional
         Exclude overlapping speech regions when extracting embeddings.
diff --git a/pyannote/audio/pipelines/speech_separation.py b/pyannote/audio/pipelines/speech_separation.py
index 28986e4e6..4172601c3 100644
--- a/pyannote/audio/pipelines/speech_separation.py
+++ b/pyannote/audio/pipelines/speech_separation.py
@@ -70,7 +70,7 @@ class SpeechSeparation(SpeakerDiarizationMixin, Pipeline):
         `segmentation_step` controls the step of this window, provided as a ratio of its
         duration. Defaults to 0.1 (i.e. 90% overlap between two consecuive windows).
     embedding : Model, str, or dict, optional
-        Pretrained embedding model. Defaults to "pyannote/embedding@2022.07".
+        Pretrained embedding model. Defaults to "speechbrain/spkrec-ecapa-voxceleb@5c0be38".
         See pyannote.audio.pipelines.utils.get_model for supported format.
     embedding_exclude_overlap : bool, optional
         Exclude overlapping speech regions when extracting embeddings.