diff --git a/.github/workflows/pr-test.yml b/.github/workflows/pr-test.yml
index beaa89acc7db..7f3971645612 100644
--- a/.github/workflows/pr-test.yml
+++ b/.github/workflows/pr-test.yml
@@ -1146,6 +1146,32 @@ jobs:
           cd test/srt
           python3 run_suite.py --suite per-commit-4-gpu-gb200 --auto-partition-id 0 --auto-partition-size 1 --timeout-per-file 3600
 
+  unit-test-backend-8-gpu-b200:
+    needs: [check-changes, call-gate, unit-test-backend-2-gpu]
+    if: |
+      (inputs.target_stage == 'unit-test-backend-8-gpu-b200') ||
+      (
+        always() &&
+        (github.event_name == 'schedule' || (!failure() && !cancelled())) &&
+        ((needs.check-changes.outputs.main_package == 'true') || (needs.check-changes.outputs.sgl_kernel == 'true'))
+      )
+    runs-on: 8-gpu-b200
+    env:
+      RUNNER_LABELS: 8-gpu-b200
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Install dependencies
+        run: |
+          IS_BLACKWELL=1 bash scripts/ci/ci_install_dependency.sh
+
+      - name: Run test
+        timeout-minutes: 45
+        run: |
+          cd test/srt
+          python3 run_suite.py --suite per-commit-8-gpu-b200 --timeout-per-file 1800
+
   pr-test-finish:
     needs:
       [
@@ -1178,6 +1204,7 @@ jobs:
         unit-test-deepep-8-gpu,
         unit-test-backend-4-gpu-b200,
         unit-test-backend-4-gpu-gb200,
+        unit-test-backend-8-gpu-b200,
       ]
     if: always()
     runs-on: ubuntu-latest
diff --git a/test/srt/run_suite.py b/test/srt/run_suite.py
index 5c899a6c24fe..e17c25fcdae6 100644
--- a/test/srt/run_suite.py
+++ b/test/srt/run_suite.py
@@ -180,7 +180,9 @@
         # TODO: Add it back after the bug is fixed
         # TestFile("test_eagle_infer_beta_dp_attention.py", 200),
     ],
-    "per-commit-8-gpu-b200": [],
+    "per-commit-8-gpu-b200": [
+        TestFile("test_mistral_large3_basic.py", 275),
+    ],
     "per-commit-4-gpu-gb200": [
         TestFile("test_cutedsl_moe.py", 300),
         TestFile("test_deepseek_v3_cutedsl_4gpu.py", 590),
diff --git a/test/srt/test_mistral_large3_basic.py b/test/srt/test_mistral_large3_basic.py
new file mode 100644
index 000000000000..3b173f1abd7c
--- /dev/null
+++ b/test/srt/test_mistral_large3_basic.py
@@ -0,0 +1,87 @@
+import os
+import unittest
+from types import SimpleNamespace
+
+from sglang.srt.utils import kill_process_tree
+from sglang.test.few_shot_gsm8k import run_eval as run_eval_few_shot_gsm8k
+from sglang.test.send_one import BenchArgs, send_one_prompt
+from sglang.test.test_utils import (
+    DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
+    DEFAULT_URL_FOR_TEST,
+    CustomTestCase,
+    is_in_ci,
+    popen_launch_server,
+    write_github_step_summary,
+)
+
+MISTRAL_LARGE3_MODEL_PATH = "mistralai/Mistral-Large-3-675B-Instruct-2512"
+
+
+class TestMistralLarge3Basic(CustomTestCase):
+    @classmethod
+    def setUpClass(cls):
+        # Set environment variable to disable JIT DeepGemm
+        os.environ["SGLANG_ENABLE_JIT_DEEPGEMM"] = "0"
+
+        cls.model = MISTRAL_LARGE3_MODEL_PATH
+        cls.base_url = DEFAULT_URL_FOR_TEST
+        other_args = [
+            "--tp",
+            "8",
+            "--attention-backend",
+            "trtllm_mla",
+            "--model-loader-extra-config",
+            '{"enable_multithread_load": true}',
+            "--chat-template",
+            "mistral",
+        ]
+        cls.process = popen_launch_server(
+            cls.model,
+            cls.base_url,
+            timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH * 5,
+            other_args=other_args,
+        )
+
+    @classmethod
+    def tearDownClass(cls):
+        kill_process_tree(cls.process.pid)
+        # Clean up environment variable
+        if "SGLANG_ENABLE_JIT_DEEPGEMM" in os.environ:
+            del os.environ["SGLANG_ENABLE_JIT_DEEPGEMM"]
+
+    def test_a_gsm8k(
+        self,
+    ):  # Append an "a" to make this test run first (alphabetically) to warm up the server
+        args = SimpleNamespace(
+            num_shots=8,
+            data_path=None,
+            num_questions=1400,
+            parallel=1400,
+            max_new_tokens=512,
+            host="http://127.0.0.1",
+            port=int(self.base_url.split(":")[-1]),
+        )
+        metrics = run_eval_few_shot_gsm8k(args)
+        print(f"{metrics=}")
+
+        if is_in_ci():
+            write_github_step_summary(
+                f"### test_gsm8k (mistral-large-3)\n" f'{metrics["accuracy"]=:.3f}\n'
+            )
+            self.assertGreater(metrics["accuracy"], 0.90)
+
+    def test_bs_1_speed(self):
+        args = BenchArgs(port=int(self.base_url.split(":")[-1]), max_new_tokens=2048)
+        acc_length, speed = send_one_prompt(args)
+
+        print(f"{speed=:.2f}")
+
+        if is_in_ci():
+            write_github_step_summary(
+                f"### test_bs_1_speed (mistral-large-3)\n" f"{speed=:.2f} token/s\n"
+            )
+            self.assertGreater(speed, 50)
+
+
+if __name__ == "__main__":
+    unittest.main()