diff --git a/.buildkite/test_areas/benchmarks.yaml b/.buildkite/test_areas/benchmarks.yaml
index 57080c46f..a30ec60ea 100644
--- a/.buildkite/test_areas/benchmarks.yaml
+++ b/.buildkite/test_areas/benchmarks.yaml
@@ -22,9 +22,10 @@ steps:
   device: b200
   num_gpus: 2
   optional: true
+  working_dir: "/vllm-workspace/"
   timeout_in_minutes: 10
   source_file_dependencies:
   - benchmarks/attention_benchmarks/
   - vllm/v1/attention/
   commands:
-  - python benchmarks/attention_benchmarks/benchmark.py --backends flash flashinfer --batch-specs "8q1s1k" --repeats 1 --warmup-iters 1
+  - python3 benchmarks/attention_benchmarks/benchmark.py --backends flash flashinfer --batch-specs "8q1s1k" --repeats 1 --warmup-iters 1