diff --git a/src/locales/en-US/benchmark.ts b/src/locales/en-US/benchmark.ts
index 5848c161..8be5e6c6 100644
--- a/src/locales/en-US/benchmark.ts
+++ b/src/locales/en-US/benchmark.ts
@@ -91,5 +91,5 @@ export default {
   'benchmark.detail.result.basic': 'Basic',
   'benchmark.form.profile.ShareGPT': 'ShareGPT',
   'benchmark.form.profile.ShareGPT.tips':
-    'Max throughput with realistic prompt length. The most commonly used baseline for GPU and model comparison.'
+    'Measures maximum throughput using real conversational data. Suitable for GPU and model performance benchmarking.'
 };
diff --git a/src/locales/ja-JP/benchmark.ts b/src/locales/ja-JP/benchmark.ts
index 5848c161..8be5e6c6 100644
--- a/src/locales/ja-JP/benchmark.ts
+++ b/src/locales/ja-JP/benchmark.ts
@@ -91,5 +91,5 @@ export default {
   'benchmark.detail.result.basic': 'Basic',
   'benchmark.form.profile.ShareGPT': 'ShareGPT',
   'benchmark.form.profile.ShareGPT.tips':
-    'Max throughput with realistic prompt length. The most commonly used baseline for GPU and model comparison.'
+    'Measures maximum throughput using real conversational data. Suitable for GPU and model performance benchmarking.'
 };
diff --git a/src/locales/ru-RU/benchmark.ts b/src/locales/ru-RU/benchmark.ts
index 5848c161..8be5e6c6 100644
--- a/src/locales/ru-RU/benchmark.ts
+++ b/src/locales/ru-RU/benchmark.ts
@@ -91,5 +91,5 @@ export default {
   'benchmark.detail.result.basic': 'Basic',
   'benchmark.form.profile.ShareGPT': 'ShareGPT',
   'benchmark.form.profile.ShareGPT.tips':
-    'Max throughput with realistic prompt length. The most commonly used baseline for GPU and model comparison.'
+    'Measures maximum throughput using real conversational data. Suitable for GPU and model performance benchmarking.'
 };
diff --git a/src/locales/zh-CN/benchmark.ts b/src/locales/zh-CN/benchmark.ts
index 1bf82060..963f98cb 100644
--- a/src/locales/zh-CN/benchmark.ts
+++ b/src/locales/zh-CN/benchmark.ts
@@ -90,5 +90,5 @@ export default {
   'benchmark.detail.result.basic': '基础信息',
   'benchmark.form.profile.ShareGPT': 'ShareGPT',
   'benchmark.form.profile.ShareGPT.tips':
-    '基于真实提示词长度的最大吞吐表现，是最常用的 GPU 与模型对比基准。'
+    '用于测量基于真实对话数据的最大吞吐量，适合进行 GPU 与模型性能基准测试。'
 };