diff --git a/.github/workflows/gpt-engineer.yml b/.github/workflows/gpt-engineer.yml
index a39165482..833026e8e 100644
--- a/.github/workflows/gpt-engineer.yml
+++ b/.github/workflows/gpt-engineer.yml
@@ -5,6 +5,8 @@ on:
     branches: [master]
   push:
     branches: [stable, master, ci-test*]
+  pull_request:
+    branches: [stable, master, ci-test*]
 
 jobs:
   regression-tests:
@@ -52,14 +54,30 @@ jobs:
           poetry install --only main
           poetry build
 
-      - name: Run regression tests
+      - name: Run regression tests (push)
+        if: ${{ github.event_name != 'pull_request' }}
         run: |
           cd agent/gpt-engineer
           make install
           source venv/bin/activate
-          pip install ../../dist/agbenchmark-0.1.0-py3-none-any.whl
-          agbenchmark start --maintain
+          pip install ../../dist/*.whl
+          
+          if [ "${GITHUB_EVENT_NAME}" != "pull_request" ]; then
+            agbenchmark start --maintain
+          else
+            agbenchmark start --maintain --mock
+            agbenchmark start --improve --mock
+            agbenchmark start --mock
+            agbenchmark start --mock --category=retrieval
+            agbenchmark start --mock --category=regression
+            agbenchmark start --mock --category=interface
+            agbenchmark start --mock --category=code
+            agbenchmark start --mock --category=memory
+            agbenchmark start --mock --category=memory --category=code
+          fi
+
         env:
+          GITHUB_EVENT_NAME: ${{ github.event_name }}
           OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
 
       - name: Upload logs as artifact
diff --git a/agbenchmark/start_benchmark.py b/agbenchmark/start_benchmark.py
index f78e86a1c..68c7932be 100644
--- a/agbenchmark/start_benchmark.py
+++ b/agbenchmark/start_benchmark.py
@@ -89,6 +89,9 @@ def start(category: str, maintain: bool, improve: bool, mock: bool) -> int:
     if mock:
         pytest_args.append("--mock")
 
+    # when used as a library, the pytest directory to execute is in the CURRENT_DIRECTORY
+    pytest_args.append(str(CURRENT_DIRECTORY))
+
     return sys.exit(pytest.main(pytest_args))
 
 
diff --git a/agent/Auto-GPT b/agent/Auto-GPT
index 5a36e43b7..cec424ad2 160000
--- a/agent/Auto-GPT
+++ b/agent/Auto-GPT
@@ -1 +1 @@
-Subproject commit 5a36e43b782fdaef8a7270109f8347f0323211d2
+Subproject commit cec424ad2504020a830c3af9f74536a420545931
diff --git a/agent/gpt-engineer b/agent/gpt-engineer
index fd705f89a..4af8c137e 160000
--- a/agent/gpt-engineer
+++ b/agent/gpt-engineer
@@ -1 +1 @@
-Subproject commit fd705f89afd53469c91935a9cae7b92a564025eb
+Subproject commit 4af8c137e82cc51fdd31c23327ceffd64194b984