diff --git a/CHANGELOG.md b/CHANGELOG.md
index 47bd81aaf91510fad8508c8c93520547fc58cee5..5a2f82aa4f2cd87f3c9ec8ab95207e66de663c28 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -28,17 +28,30 @@ fixed.  The version numbering adheres to the
 
 * Extended operator support:
   * Ethos-U85:
-    * TFLite: LOG and RELU_0_TO_1
+    * TFLite:
+      * LOG, LSTM, STRIDED_SLICE with strides larger than one, RELU_0_TO_1
   * Ethos-U55/Ethos-U65:
-    * TFLite: RELU_0_TO_1
+    * TFLite:
+      * RELU_0_TO_1
+* Upgrade TensorFlow Lite support to version 2.18.1
 
 **Interface changes:**
 
-  * None
+* None
 
 **Reported defect fixes:**
 
+* Ethos-U55/Ethos-U65
   * None
+* Ethos-U85
+  * Improved weight format selection for models with 2:4 sparse weights (MLCE-1473)
+* General
+  * Fixed issue with slicing of 64-bit bias causing segmentation fault (MLCE-1476)
+  * Fixed graph connection issues caused by Reshape removal (MLCE-1496)
+  * Fixed unintended disabling of the `--cpu-tensor-alignment` option (MLCE-1472)
+  * Fixed issue where fused activations are recorded as CPU operators (MLCE-1461)
+  * Fixed various issues with MAC count in summary (MLCE-1493 & MLCE-1461)
+
 
 ## Release 4.2.0 - 27/02/2025
 
diff --git a/TESTING.md b/TESTING.md
index 016355a04cbe3447c015b7f7db118f4df2bc737c..f76f43be027e8bc69033c4596f06a2fd67ae1b0c 100644
--- a/TESTING.md
+++ b/TESTING.md
@@ -36,7 +36,8 @@ kernel).  The following list indicates which TensorFlow version a particular
 Vela version was tested against.  It also indicates which version of the
 TensorFlow Lite/LiteRT flatbuffer schema was supported.
 
-* Vela 4.0.0 to current supports TensorFlow 2.17
+* Vela 4.3.0 to current supports TensorFlow 2.18.1
+* Vela 4.0.0 to 4.2.0 supports TensorFlow 2.17
 * Vela 3.12.0 supports TensorFlow 2.16
 * Vela 3.11.0 supports TensorFlow 2.15
 * Vela 3.10.0 supports TensorFlow 2.14