diff --git a/frontend/src/i18n/locales/en.ts b/frontend/src/i18n/locales/en.ts
index 54185129..d6c87d52 100644
--- a/frontend/src/i18n/locales/en.ts
+++ b/frontend/src/i18n/locales/en.ts
@@ -1669,9 +1669,23 @@ export default {
         title: 'OpenAI Messages Dispatch',
         allowDispatch: 'Allow /v1/messages dispatch',
         allowDispatchHint: 'When enabled, API keys in this OpenAI group can dispatch requests through /v1/messages endpoint',
-        defaultModel: 'Default mapped model',
-        defaultModelPlaceholder: 'e.g., gpt-4.1',
-        defaultModelHint: 'When account has no model mapping configured, all request models will be mapped to this model'
+        familyMappingTitle: 'Family Default Mapping',
+        familyMappingHint: 'Requests that match the Opus, Sonnet, or Haiku families will prefer the target model configured here.',
+        opusModel: 'Opus Target Model',
+        opusModelPlaceholder: 'e.g., gpt-5.4',
+        sonnetModel: 'Sonnet Target Model',
+        sonnetModelPlaceholder: 'e.g., gpt-5.3-codex',
+        haikuModel: 'Haiku Target Model',
+        haikuModelPlaceholder: 'e.g., gpt-5.4-mini',
+        exactMappingTitle: 'Exact Model Overrides',
+        exactMappingHint: 'Exact Claude model overrides take priority over the family defaults and can route a specific Claude model to a different target model.',
+        noExactMappings: 'No exact model overrides yet',
+        addExactMapping: 'Add Exact Mapping',
+        claudeModel: 'Claude Model',
+        claudeModelPlaceholder: 'e.g., claude-sonnet-4-5-20250929',
+        targetModel: 'Target Model',
+        targetModelPlaceholder: 'e.g., gpt-5.4',
+        removeExactMapping: 'Remove Exact Mapping'
       },
       invalidRequestFallback: {
         title: 'Invalid Request Fallback Group',
diff --git a/frontend/src/i18n/locales/zh.ts b/frontend/src/i18n/locales/zh.ts
index b16dc715..2038970a 100644
--- a/frontend/src/i18n/locales/zh.ts
+++ b/frontend/src/i18n/locales/zh.ts
@@ -1755,9 +1755,23 @@ export default {
         title: 'OpenAI Messages 调度配置',
         allowDispatch: '允许 /v1/messages 调度',
         allowDispatchHint: '启用后，此 OpenAI 分组的 API Key 可以通过 /v1/messages 端点调度请求',
-        defaultModel: '默认映射模型',
-        defaultModelPlaceholder: '例如: gpt-4.1',
-        defaultModelHint: '当账号未配置模型映射时，所有请求模型将映射到此模型'
+        familyMappingTitle: '系列默认映射',
+        familyMappingHint: '当请求命中 Opus、Sonnet、Haiku 系列时，会优先使用这里配置的目标模型。',
+        opusModel: 'Opus 映射模型',
+        opusModelPlaceholder: '例如: gpt-5.4',
+        sonnetModel: 'Sonnet 映射模型',
+        sonnetModelPlaceholder: '例如: gpt-5.3-codex',
+        haikuModel: 'Haiku 映射模型',
+        haikuModelPlaceholder: '例如: gpt-5.4-mini',
+        exactMappingTitle: '精确模型覆盖',
+        exactMappingHint: '精确 Claude 模型覆盖优先级高于系列默认映射，可将某个具体 Claude 模型单独映射到不同的目标模型。',
+        noExactMappings: '暂无精确模型覆盖规则',
+        addExactMapping: '添加精确映射',
+        claudeModel: 'Claude 模型',
+        claudeModelPlaceholder: '例如: claude-sonnet-4-5-20250929',
+        targetModel: '目标模型',
+        targetModelPlaceholder: '例如: gpt-5.4',
+        removeExactMapping: '删除精确映射'
       },
       invalidRequestFallback: {
         title: '无效请求兜底分组',