diff --git a/model_server/poetry.lock b/model_server/poetry.lock
index ce40bac9..c5966dba 100644
--- a/model_server/poetry.lock
+++ b/model_server/poetry.lock
@@ -2843,6 +2843,23 @@ pytest = ">=8.2,<9"
 docs = ["sphinx (>=5.3)", "sphinx-rtd-theme (>=1.0)"]
 testing = ["coverage (>=6.2)", "hypothesis (>=5.7.1)"]
 
+[[package]]
+name = "pytest-retry"
+version = "1.6.3"
+description = "Adds the ability to retry flaky tests in CI environments"
+optional = false
+python-versions = ">=3.9"
+files = [
+    {file = "pytest_retry-1.6.3-py3-none-any.whl", hash = "sha256:e96f7df77ee70b0838d1085f9c3b8b5b7d74bf8947a0baf32e2b8c71b27683c8"},
+    {file = "pytest_retry-1.6.3.tar.gz", hash = "sha256:36ccfa11c8c8f9ddad5e20375182146d040c20c4a791745139c5a99ddf1b557d"},
+]
+
+[package.dependencies]
+pytest = ">=7.0.0"
+
+[package.extras]
+dev = ["black", "flake8", "isort", "mypy"]
+
 [[package]]
 name = "python-dateutil"
 version = "2.9.0.post0"
@@ -4318,4 +4335,4 @@ type = ["pytest-mypy"]
 [metadata]
 lock-version = "2.0"
 python-versions = ">=3.12"
-content-hash = "39638c5ce345e597b8cd12b4ec5efb67e745492b7d4e05bd576793a6352a9e3a"
+content-hash = "0a1026a25a578b3cc049ebe26691093b9a2d8ec0ad5cbec57e9949a138c177b8"
diff --git a/model_server/pyproject.toml b/model_server/pyproject.toml
index aef566a0..4f212a8e 100644
--- a/model_server/pyproject.toml
+++ b/model_server/pyproject.toml
@@ -35,6 +35,7 @@ opentelemetry-sdk = "^1.28.0"
 opentelemetry-exporter-otlp = "^1.28.0"
 opentelemetry-instrumentation-fastapi = "^0.49b0"
 overrides = "^7.7.0"
+pytest-retry = "^1.6.3"
 
 [tool.poetry.scripts]
 archgw_modelserver = "src.cli:run_server"
@@ -46,6 +47,6 @@ build-backend = "poetry.core.masonry.api"
 [tool.pytest.ini_options]
 python_files = ["test*.py"]
 addopts = ["-v", "-s"]
-retries = 5
+retries = 2
 retry_delay = 0.5
 cumulative_timing = false