diff --git a/README.md b/README.md
index 9ac15f41..3c5700a0 100644
--- a/README.md
+++ b/README.md
@@ -101,16 +101,15 @@ Starting from CUTLASS 3.0, CUTLASS removed support for the following:
# Performance
-
+
+
CUTLASS primitives are very efficient. When used to construct device-wide GEMM kernels,
they exhibit peak performance comparable to cuBLAS for scalar GEMM
-computations. The above figure shows CUTLASS performance relative to cuBLAS
-for large matrix dimensions on an [NVIDIA H100](https://www.nvidia.com/en-us/data-center/h100/) (NVIDIA Hopper architecture),
-an [NVIDIA L40](https://www.nvidia.com/en-us/data-center/l40/) (NVIDIA Ada architecture),
-an [NVIDIA A100](https://www.nvidia.com/en-us/data-center/a100/) (NVIDIA Ampere architecture),
-and an [NVIDIA A40](https://www.nvidia.com/en-us/data-center/a40/) (NVIDIA Ampere architecture).
-CUTLASS 3.0 was compiled with the [CUDA 12.0 Toolkit](https://developer.nvidia.com/cuda-downloads).
+computations. The above figure shows the continual CUTLASS performance improvements
+on an [NVIDIA H100](https://www.nvidia.com/en-us/data-center/h100/) (NVIDIA Hopper architecture) since
+CUTLASS 3.1.
+CUTLASS 3.5.1 was compiled with the [CUDA 12.5u1 Toolkit](https://developer.nvidia.com/cuda-downloads).
Tensor Core operations are implemented using CUDA's
[mma](https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#warp-level-matrix-instructions-mma) and
[wgmma](https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#asynchronous-warpgroup-level-matrix-instructions) instructions.
diff --git a/media/images/cutlass-3.5.1-gemm-peak-performance-fp8.png b/media/images/cutlass-3.5.1-gemm-peak-performance-fp8.png
new file mode 100644
index 00000000..bca203c0
Binary files /dev/null and b/media/images/cutlass-3.5.1-gemm-peak-performance-fp8.png differ
diff --git a/media/images/cutlass-3.5.1-gemm-peak-performance.png b/media/images/cutlass-3.5.1-gemm-peak-performance.png
new file mode 100644
index 00000000..90b0bd6d
Binary files /dev/null and b/media/images/cutlass-3.5.1-gemm-peak-performance.png differ