Revert "tiny remove deprecated endpoint call"

docs/advanced_features/router.md

@@ -45,7 +45,7 @@ SGLang Model Gateway is a high-performance model-routing gateway for large-scale
 ## Architecture
 
 ### Control Plane
-- **Worker Manager** discovers capabilities (`/server_info`, `/model_info`), tracks load, and registers/removes workers in the shared registry.
+- **Worker Manager** discovers capabilities (`/get_server_info`, `/get_model_info`), tracks load, and registers/removes workers in the shared registry.
 - **Job Queue** serializes add/remove requests and exposes status (`/workers/{url}`) so clients can track onboarding progress.
 - **Load Monitor** feeds cache-aware and power-of-two policies with live worker load statistics.
 - **Health Checker** continuously probes workers and updates readiness, circuit breaker state, and router metrics.

docs/advanced_features/server_arguments.md

@@ -219,7 +219,7 @@ Please consult the documentation below and [server_args.py](https://github.com/s
 | Argument | Description | Defaults | Options |
 | --- | --- | --- | --- |
 | `--json-model-override-args` | A dictionary in JSON string format used to override default model configurations. | `{}` | Type: str |
-| `--preferred-sampling-params` | json-formatted sampling settings that will be returned in /model_info | `None` | Type: str |
+| `--preferred-sampling-params` | json-formatted sampling settings that will be returned in /get_model_info | `None` | Type: str |
 
 ## LoRA
 | Argument | Description | Defaults | Options |

docs/basic_usage/native_api.ipynb

@@ -9,8 +9,8 @@
     "Apart from the OpenAI compatible APIs, the SGLang Runtime also provides its native server APIs. We introduce the following APIs:\n",
     "\n",
     "- `/generate` (text generation model)\n",
-    "- `/model_info`\n",
-    "- `/server_info`\n",
+    "- `/get_model_info`\n",
+    "- `/get_server_info`\n",
     "- `/health`\n",
     "- `/health_generate`\n",
     "- `/flush_cache`\n",
@@ -99,7 +99,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "url = f\"http://localhost:{port}/model_info\"\n",
+    "url = f\"http://localhost:{port}/get_model_info\"\n",
     "\n",
     "response = requests.get(url)\n",
     "response_json = response.json()\n",
@@ -127,7 +127,7 @@
    "source": [
     "## Get Server Info\n",
     "Gets the server information including CLI arguments, token limits, and memory pool sizes.\n",
-    "- Note: `server_info` merges the following deprecated endpoints:\n",
+    "- Note: `get_server_info` merges the following deprecated endpoints:\n",
     "  - `get_server_args`\n",
     "  - `get_memory_pool_size` \n",
     "  - `get_max_total_num_tokens`"
@@ -139,7 +139,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "url = f\"http://localhost:{port}/server_info\"\n",
+    "url = f\"http://localhost:{port}/get_server_info\"\n",
     "\n",
     "response = requests.get(url)\n",
     "print_highlight(response.text)"

docs/developer_guide/bench_serving.md

@@ -352,4 +352,4 @@ python3 -m sglang.bench_serving \
 ### Notes
 
 - The script raises the file descriptor soft limit (`RLIMIT_NOFILE`) to help with many concurrent connections.
-- For sglang, `/server_info` is queried post-run to report speculative decoding accept length when available.
+- For sglang, `/get_server_info` is queried post-run to report speculative decoding accept length when available.

python/sglang/bench_one_batch_server.py

@@ -339,7 +339,7 @@ def run_one_case(
     output_throughput = batch_size * output_len / (latency - last_ttft)
     overall_throughput = batch_size * (input_len + output_len) / latency
 
-    server_info = requests.get(url + "/server_info").json()
+    server_info = requests.get(url + "/get_server_info").json()
     internal_state = server_info.get("internal_states", [{}])
     last_gen_throughput = internal_state[0].get("last_gen_throughput", None) or -1
     acc_length = internal_state[0].get("avg_spec_accept_length", None) or -1
@@ -451,7 +451,7 @@ def run_benchmark(server_args: ServerArgs, bench_args: BenchArgs):
         proc, base_url = launch_server_process(server_args)
 
     # Get tokenizer
-    server_info = requests.get(base_url + "/server_info").json()
+    server_info = requests.get(base_url + "/get_server_info").json()
     if "tokenizer_path" in server_info:
         tokenizer_path = server_info["tokenizer_path"]
     elif "prefill" in server_info:

python/sglang/bench_serving.py

@@ -2062,7 +2062,7 @@ async def limited_request_func(request_func_input, pbar):
 
     if "sglang" in backend:
         server_info = requests.get(
-            base_url + "/server_info", headers=get_auth_headers()
+            base_url + "/get_server_info", headers=get_auth_headers()
         )
         if server_info.status_code == 200:
             server_info_json = server_info.json()
@@ -2180,7 +2180,7 @@ async def limited_request_func(request_func_input, pbar):
     print("{:<40} {:<10.2f}".format("Max ITL (ms):", metrics.max_itl_ms))
     print("=" * 50)
 
-    resp = requests.get(base_url + "/server_info", headers=get_auth_headers())
+    resp = requests.get(base_url + "/get_server_info", headers=get_auth_headers())
     server_info = resp.json() if resp.status_code == 200 else None
 
     if (

python/sglang/lang/backend/runtime_endpoint.py

@@ -38,7 +38,7 @@ def __init__(
         self.verify = verify
 
         res = http_request(
-            self.base_url + "/model_info",
+            self.base_url + "/get_model_info",
             api_key=self.api_key,
             verify=self.verify,
         )
@@ -66,7 +66,7 @@ def flush_cache(self):
 
     def get_server_info(self):
         res = http_request(
-            self.base_url + "/server_info",
+            self.base_url + "/get_server_info",
             api_key=self.api_key,
             verify=self.verify,
         )
@@ -514,7 +514,7 @@ def encode(
 
     async def get_server_info(self):
         async with aiohttp.ClientSession() as session:
-            async with session.get(f"{self.url}/server_info") as response:
+            async with session.get(f"{self.url}/get_server_info") as response:
                 if response.status == 200:
                     return await response.json()
                 else:

python/sglang/profiler.py

@@ -41,7 +41,7 @@ def run_profile(
     # Dump server args.
     file_path = Path(output_dir) / "server_args.json"
     if not file_path.exists():
-        response = requests.get(url + "/server_info")
+        response = requests.get(url + "/get_server_info")
         response.raise_for_status()
         server_args_data = response.json()
         with open(file_path, "w") as file:

python/sglang/srt/entrypoints/http_server.py

@@ -1480,7 +1480,7 @@ def _execute_server_warmup(
     for _ in range(120):
         time.sleep(1)
         try:
-            res = requests.get(url + "/model_info", timeout=5, headers=headers)
+            res = requests.get(url + "/get_model_info", timeout=5, headers=headers)
             assert res.status_code == 200, f"{res=}, {res.text=}"
             success = True
             break

python/sglang/srt/server_args.py

@@ -2801,7 +2801,7 @@ def add_cli_args(parser: argparse.ArgumentParser):
         parser.add_argument(
             "--preferred-sampling-params",
             type=str,
-            help="json-formatted sampling settings that will be returned in /model_info",
+            help="json-formatted sampling settings that will be returned in /get_model_info",
         )
 
         # LoRA

scripts/playground/bench_speculative.py

@@ -120,7 +120,7 @@ def send_one_batch(base_url, num_prompts, batch_size, processor, is_multimodal):
     acc_length = results["accept_length"] or 1.0
     avg_output_token = results["total_output_tokens"] / results["completed"]
 
-    server_info = requests.get(base_url + "/server_info").json()
+    server_info = requests.get(base_url + "/get_server_info").json()
     # We use 20% percentile instead of median on purpose
     step_time = np.percentile(
         server_info["internal_states"][0]["step_time_dict"][str(batch_size)], 20

test/manual/test_deepseek_v32_cp_single_node.py

@@ -79,7 +79,7 @@ def test_a_gsm8k(
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/server_info")
+        server_info = requests.get(self.base_url + "/get_server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/manual/test_eagle_infer_beta_dp_attention.py

@@ -77,7 +77,7 @@ def test_a_gsm8k(
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/server_info")
+        server_info = requests.get(self.base_url + "/get_server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/manual/test_weight_version.py

@@ -3,7 +3,7 @@
 
 This test suite verifies the weight_version feature implementation including:
 1. Default weight_version setting
-2. /weight_version endpoint
+2. /get_weight_version endpoint
 3. /update_weight_version endpoint
 4. /generate request meta_info contains weight_version
 5. OpenAI API response metadata contains weight_version
@@ -48,13 +48,13 @@ def tearDownClass(cls):
     def test_weight_version_comprehensive(self):
         """Comprehensive test for all weight_version functionality."""
 
-        response = requests.get(f"{self.base_url}/model_info")
+        response = requests.get(f"{self.base_url}/get_model_info")
         self.assertEqual(response.status_code, 200)
         data = response.json()
         self.assertIn("weight_version", data)
         self.assertEqual(data["weight_version"], "test_version_1.0")
 
-        response = requests.get(f"{self.base_url}/weight_version")
+        response = requests.get(f"{self.base_url}/get_weight_version")
         self.assertEqual(response.status_code, 200)
         data = response.json()
         self.assertIn("weight_version", data)
@@ -114,7 +114,7 @@ def test_weight_version_comprehensive(self):
         self.assertTrue(data["success"])
         self.assertEqual(data["new_version"], "updated_version_2.0")
 
-        response = requests.get(f"{self.base_url}/weight_version")
+        response = requests.get(f"{self.base_url}/get_weight_version")
         self.assertEqual(response.status_code, 200)
         data = response.json()
         self.assertEqual(data["weight_version"], "updated_version_2.0")
@@ -148,13 +148,13 @@ def test_weight_version_comprehensive(self):
         self.assertTrue(data["success"])
         self.assertEqual(data["new_version"], "final_version_3.0")
 
-        # Check /weight_version
-        response = requests.get(f"{self.base_url}/weight_version")
+        # Check /get_weight_version
+        response = requests.get(f"{self.base_url}/get_weight_version")
         self.assertEqual(response.status_code, 200)
         self.assertEqual(response.json()["weight_version"], "final_version_3.0")
 
-        # Check /model_info
-        response = requests.get(f"{self.base_url}/model_info")
+        # Check /get_model_info
+        response = requests.get(f"{self.base_url}/get_model_info")
         self.assertEqual(response.status_code, 200)
         self.assertEqual(response.json()["weight_version"], "final_version_3.0")
 
@@ -193,7 +193,7 @@ def test_update_weight_version_with_weight_updates(self):
         print("Testing weight_version update with real weight operations...")
 
         # Get current model info for reference
-        model_info_response = requests.get(f"{self.base_url}/model_info")
+        model_info_response = requests.get(f"{self.base_url}/get_model_info")
         self.assertEqual(model_info_response.status_code, 200)
         current_model_path = model_info_response.json()["model_path"]
 
@@ -214,7 +214,7 @@ def test_update_weight_version_with_weight_updates(self):
         )
 
         # Verify version was updated
-        version_response = requests.get(f"{self.base_url}/weight_version")
+        version_response = requests.get(f"{self.base_url}/get_weight_version")
         self.assertEqual(version_response.status_code, 200)
         self.assertEqual(
             version_response.json()["weight_version"], "disk_update_v2.0.0"

test/srt/ep/test_deepep_large.py

@@ -143,7 +143,7 @@ def test_gsm8k(self):
 
         self.assertGreater(metrics["accuracy"], 0.92)
 
-        server_info = requests.get(self.base_url + "/server_info")
+        server_info = requests.get(self.base_url + "/get_server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/srt/ep/test_deepep_small.py

@@ -307,7 +307,7 @@ def test_gsm8k(self):
 
         self.assertGreater(metrics["accuracy"], 0.60)
 
-        server_info = requests.get(self.base_url + "/server_info")
+        server_info = requests.get(self.base_url + "/get_server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]
@@ -381,7 +381,7 @@ def test_gsm8k(self):
 
         self.assertGreater(metrics["accuracy"], 0.60)
 
-        server_info = requests.get(self.base_url + "/server_info")
+        server_info = requests.get(self.base_url + "/get_server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/srt/hicache/test_hicache_variants.py

@@ -152,7 +152,7 @@ def test_mmlu(self):
         self.assertGreaterEqual(metrics["score"], self.expected_mmlu_score)
 
         # EAGLE-specific check
-        server_info = requests.get(self.base_url + "/server_info")
+        server_info = requests.get(self.base_url + "/get_server_info")
         print(f"{server_info=}")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"

test/srt/quant/test_w4a8_deepseek_v3.py

@@ -103,7 +103,7 @@ def test_gsm8k(
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/server_info")
+        server_info = requests.get(self.base_url + "/get_server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/srt/rl/test_update_weights_from_disk.py

@@ -113,7 +113,7 @@ def run_decode_random(self, max_new_tokens=32):
         return response.json()
 
     def get_model_info(self):
-        response = requests.get(self.base_url + "/model_info")
+        response = requests.get(self.base_url + "/get_model_info")
         model_path = response.json()["model_path"]
         print(json.dumps(response.json()))
         return model_path
@@ -254,7 +254,7 @@ def run_decode(self, max_new_tokens=32):
         return response.json()
 
     def get_model_info(self):
-        response = requests.get(self.base_url + "/model_info")
+        response = requests.get(self.base_url + "/get_model_info")
         model_path = response.json()["model_path"]
         print(json.dumps(response.json()))
         return model_path
@@ -386,7 +386,7 @@ def run_decode():
             return response.json()["text"]
 
         def get_model_info():
-            response = requests.get(base_url + "/model_info")
+            response = requests.get(base_url + "/get_model_info")
             model_path = response.json()["model_path"]
             print(json.dumps(response.json()))
             return model_path

test/srt/rl/test_update_weights_from_tensor.py

@@ -209,7 +209,7 @@ def run_decode(self, max_new_tokens=32):
         return response.json()
 
     def get_model_info(self):
-        response = requests.get(self.base_url + "/model_info")
+        response = requests.get(self.base_url + "/get_model_info")
         model_path = response.json()["model_path"]
         print(json.dumps(response.json()))
         return model_path

test/srt/test_data_parallelism.py

@@ -65,12 +65,12 @@ def test_update_weight(self):
 
     def test_get_memory_pool_size(self):
         # use `get_server_info` instead since `get_memory_pool_size` is merged into `get_server_info`
-        response = requests.get(self.base_url + "/server_info")
+        response = requests.get(self.base_url + "/get_server_info")
         assert response.status_code == 200
 
         time.sleep(1)
 
-        response = requests.get(self.base_url + "/server_info")
+        response = requests.get(self.base_url + "/get_server_info")
         assert response.status_code == 200
 
 

test/srt/test_deepseek_v32_mtp.py

@@ -69,7 +69,7 @@ def test_a_gsm8k(
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/server_info")
+        server_info = requests.get(self.base_url + "/get_server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/srt/test_deepseek_v3_fp4_4gpu.py

@@ -143,7 +143,7 @@ def test_a_gsm8k(
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/server_info")
+        server_info = requests.get(self.base_url + "/get_server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/srt/test_deepseek_v3_mtp.py

@@ -67,7 +67,7 @@ def test_a_gsm8k(
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/server_info")
+        server_info = requests.get(self.base_url + "/get_server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/srt/test_dp_attention.py

@@ -112,7 +112,7 @@ def test_gsm8k(self):
 
         self.assertGreater(metrics["accuracy"], 0.60)
 
-        server_info = requests.get(self.base_url + "/server_info")
+        server_info = requests.get(self.base_url + "/get_server_info")
         avg_spec_accept_length = server_info.json()["internal_states"][0][
             "avg_spec_accept_length"
         ]

test/srt/test_eagle_dp_attention.py

@@ -78,7 +78,7 @@ def test_a_gsm8k(self):
         metrics = run_eval_few_shot_gsm8k(args)
         print(f"{metrics=}")
 
-        server_info = requests.get(self.base_url + "/server_info")
+        server_info = requests.get(self.base_url + "/get_server_info")
         server_data = server_info.json()
 
         # Try to get avg_spec_accept_length

test/srt/test_eagle_infer_b.py

@@ -144,7 +144,7 @@ def test_gsm8k(self):
         print(f"{metrics=}")
         self.assertGreater(metrics["accuracy"], 0.20)
 
-        server_info = requests.get(self.base_url + "/server_info").json()
+        server_info = requests.get(self.base_url + "/get_server_info").json()
         avg_spec_accept_length = server_info["internal_states"][0][
             "avg_spec_accept_length"
         ]