diff --git a/fedot/core/operations/evaluation/operation_implementations/models/tabpfn.py b/fedot/core/operations/evaluation/operation_implementations/models/tabpfn.py
index 3341a0b468..1494b49e20 100644
--- a/fedot/core/operations/evaluation/operation_implementations/models/tabpfn.py
+++ b/fedot/core/operations/evaluation/operation_implementations/models/tabpfn.py
@@ -10,6 +10,16 @@
 
 
 class FedotTabPFNImplementation(ModelImplementation):
+    """FEDOT implementation of TabPFN that uses tabpfn model version 2.
+    To use the newer TabPFN v2.5 model, you must accept the non-commercial usage license
+    in your Hugging Face account at https://huggingface.co/Prior-Labs/tabpfn_2_5.
+    After accepting the license, update the `tabpfn` library to the latest version
+    (currently 6.0.6) and log in to your Hugging Face account, for example using env variable:
+    `os.environ["HF_TOKEN"] = "your_hf_token"`. The library will then automatically
+    use the v2.5 model. If authentication with a token fails, you will get an error.
+
+    """
+
     __operation_params = [
         'enable_categorical',
         'max_samples',
diff --git a/other_requirements/extra.txt b/other_requirements/extra.txt
index b0375b9847..de76164a23 100644
--- a/other_requirements/extra.txt
+++ b/other_requirements/extra.txt
@@ -1,7 +1,7 @@
 # DNNs
 tensorflow >= 2.8.0; python_version >= '3.8'
 torch >= 1.9.0
-tabpfn >= 2.0.0
+tabpfn >= 2.0.0, <= 2.2.1
 
 # Images
 opencv-python >= 4.5.5.64