diff --git a/.github/workflows/test-gpu.yml b/.github/workflows/test-gpu.yml
index c6f19a92a012..ee08b9332e4a 100644
--- a/.github/workflows/test-gpu.yml
+++ b/.github/workflows/test-gpu.yml
@@ -78,6 +78,9 @@ jobs:
           sudo rm -rf build || true
           sudo rm -rf bin || true
           sudo rm -rf dist || true
+          sudo docker logs $(sudo docker ps -q --filter ancestor=localai-tests) > logs.txt
+          sudo cat logs.txt || true
+          sudo rm -rf logs.txt
           make clean || true
           make teardown-e2e || true
           docker system prune -f -a --volumes || true
\ No newline at end of file
diff --git a/Makefile b/Makefile
index b088c28b5159..924aebd8c6a5 100644
--- a/Makefile
+++ b/Makefile
@@ -340,7 +340,7 @@ run-e2e-image:
 
 test-e2e:
 	@echo 'Running e2e tests'
-	LOCALAI_API=http://$(E2E_BRIDGE_IP):5390 $(GOCMD) run github.com/onsi/ginkgo/v2/ginkgo --flake-attempts 5 -v -r ./tests/e2e
+	LOCALAI_API=http://$(E2E_BRIDGE_IP):5390/v1 $(GOCMD) run github.com/onsi/ginkgo/v2/ginkgo --flake-attempts 5 -v -r ./tests/e2e
 
 teardown-e2e:
 	rm -rf ./tests/e2e-fixtures/ggllm-test-model.bin
diff --git a/tests/e2e/e2e_test.go b/tests/e2e/e2e_test.go
index fe3bd84c6466..ecf8ec2b4136 100644
--- a/tests/e2e/e2e_test.go
+++ b/tests/e2e/e2e_test.go
@@ -34,7 +34,7 @@ var _ = Describe("E2E test", func() {
 		// Check that the GPU was used
 		AfterEach(func() {
 			// Execute docker logs $$(docker ps -q --filter ancestor=localai-tests) as a command and check the output
-			cmd := exec.Command("/bin/bash", "-xce", "docker logs $$(docker ps -q --filter ancestor=localai-tests)")
+			cmd := exec.Command("/bin/bash", "-xce", "docker logs $(docker ps -q --filter ancestor=localai-tests)")
 			out, err := cmd.CombinedOutput()
 			Expect(err).ToNot(HaveOccurred())
 			Expect(string(out)).To(ContainSubstring("found 1 CUDA devices"))
@@ -45,7 +45,7 @@ var _ = Describe("E2E test", func() {
 			It("streams chat tokens", func() {
 				models, err := client.ListModels(context.TODO())
 				Expect(err).ToNot(HaveOccurred())
-				Expect(models.Models).ToNot(BeEmpty())
+				Expect(models.Models).ToNot(BeEmpty(), models.Models)
 
 				model := models.Models[0].ID
 				resp, err := client.CreateChatCompletion(context.TODO(), openai.ChatCompletionRequest{