diff --git a/server-rs/crates/platform-image/src/image_provider/protocol/request.rs b/server-rs/crates/platform-image/src/image_provider/protocol/request.rs
index d1d07867a..e5132386d 100644
--- a/server-rs/crates/platform-image/src/image_provider/protocol/request.rs
+++ b/server-rs/crates/platform-image/src/image_provider/protocol/request.rs
@@ -99,14 +99,19 @@ pub fn normalize_image_model(model: &str) -> &str {
     }
 }
 
-/// Resolves a persisted/concrete model to its owning provider without rewriting
-/// the model string. Legacy GPT Image 2 values remain readable, but are routed
-/// to the current Tiantoken task boundary instead of VectorEngine.
+/// Resolves a concrete provider model to its owning provider without rewriting the
+/// model string.
+///
+/// 中文注释：白名单是 provider 边界，不放业务名和已退役模型——
+/// `gpt-image-2.5-flare-c` / `gpt-image-2.5-sunburst-c` 走 Tiantoken，
+/// nanobanana 走 VectorEngine。业务模型名 `gpt-image-2.5`
+/// (`GPT_IMAGE_2_5_BUSINESS_NAME`) 必须在任务边界解析成具体 key；`gpt-image-2-c`
+/// 只是历史 fallback / 审计字符串，永远不会作为 provider 或业务 model 传入，
+/// 命中即拒绝。`gpt-image-2` 仍按历史可读值接受。
 pub fn resolve_image_provider(model: &str) -> Result<ImageProvider, &'static str> {
     match normalize_image_model(model) {
         GPT_IMAGE_2_MODEL
         | GPT_IMAGE_2_C_MODEL
-        | GPT_IMAGE_2_5_BUSINESS_NAME
         | GPT_IMAGE_2_5_GENERATION_MODEL
         | GPT_IMAGE_2_5_EDIT_MODEL => Ok(ImageProvider::Tiantoken),
         NANOBANANA_2_MODEL => Ok(ImageProvider::VectorEngine),
diff --git a/server-rs/crates/platform-image/tests/image_provider.rs b/server-rs/crates/platform-image/tests/image_provider.rs
index 893fab966..f4d84c3f4 100644
--- a/server-rs/crates/platform-image/tests/image_provider.rs
+++ b/server-rs/crates/platform-image/tests/image_provider.rs
@@ -1,10 +1,11 @@
 use platform_image::image_provider::{
-    GPT_IMAGE_2_5_EDIT_MODEL, GPT_IMAGE_2_5_GENERATION_MODEL, GPT_IMAGE_2_C_MODEL,
-    GPT_IMAGE_2_MODEL, ImageProvider, ImageProviderSettings, PlatformImageError, ReferenceImage,
-    VECTOR_ENGINE_PROVIDER, build_image_http_client, build_image_request_body,
-    build_image_request_body_with_model, build_nanobanana_generate_content_request_body,
-    create_image_edit, create_image_generation, create_nanobanana_generate_content,
-    images_edit_url, images_generation_url, nanobanana_generate_content_url,
+    GPT_IMAGE_2_5_BUSINESS_NAME, GPT_IMAGE_2_5_EDIT_MODEL, GPT_IMAGE_2_5_GENERATION_MODEL,
+    GPT_IMAGE_2_C_MODEL, GPT_IMAGE_2_MODEL, ImageProvider, ImageProviderSettings,
+    PlatformImageError, ReferenceImage, VECTOR_ENGINE_PROVIDER, build_image_http_client,
+    build_image_request_body, build_image_request_body_with_model,
+    build_nanobanana_generate_content_request_body, create_image_edit, create_image_generation,
+    create_nanobanana_generate_content, images_edit_url, images_generation_url,
+    nanobanana_generate_content_url, resolve_image_provider,
 };
 use std::{
     sync::{
@@ -227,6 +228,26 @@ fn nanobanana_generate_content_url_blank_model_falls_back_to_nanobanana_model()
     );
 }
 
+#[test]
+fn resolve_image_provider_rejects_business_model_and_accepts_concrete_models() {
+    // 中文注释：业务模型名不是 provider model，必须在 api-server 任务边界先解析成具体 model。
+    assert!(resolve_image_provider(GPT_IMAGE_2_5_BUSINESS_NAME).is_err());
+
+    assert_eq!(
+        resolve_image_provider(GPT_IMAGE_2_5_GENERATION_MODEL),
+        Ok(ImageProvider::Tiantoken)
+    );
+    assert_eq!(
+        resolve_image_provider(GPT_IMAGE_2_5_EDIT_MODEL),
+        Ok(ImageProvider::Tiantoken)
+    );
+    assert_eq!(
+        resolve_image_provider("gemini-3.1-flash-image-preview"),
+        Ok(ImageProvider::VectorEngine)
+    );
+    assert!(resolve_image_provider("unknown-model").is_err());
+}
+
 #[tokio::test]
 async fn image_edit_retries_send_timeout_once_and_succeeds() {
     let listener = TcpListener::bind("127.0.0.1:0")