diff --git a/src/shared/key-management/openai/checker.ts b/src/shared/key-management/openai/checker.ts
index 93d6c5e..1bf353d 100644
--- a/src/shared/key-management/openai/checker.ts
+++ b/src/shared/key-management/openai/checker.ts
@@ -334,11 +334,9 @@ export class OpenAIKeyChecker extends KeyCheckerBase<OpenAIKey> {
   }
 
   /**
-   * Tests whether the key has access to the gpt-image-1 model by making a small test request.
-   * Returns true if the key has access, false otherwise.
-   * 
-   * Ensures that the key can actually generate images by verifying the API response
-   * contains the expected data structure for successful image generation.
+   * Tests whether the key has access to the gpt-image-1 model by making a test request.
+   * Returns true if the API accepts the request without specific access errors.
+   * Does not wait for actual image generation to complete.
    */
   public async testGptImageAccess(key: OpenAIKey): Promise<boolean> {
     this.log.info({ key: key.hash }, "Testing gpt-image-1 access");
@@ -346,85 +344,96 @@ export class OpenAIKeyChecker extends KeyCheckerBase<OpenAIKey> {
     try {
       const payload = {
         model: "gpt-image-1",
-        prompt: "A cute baby sea otter",
+        prompt: "A simple test image",
         n: 1,
         size: "auto",
         quality: "low",
       };
       
-      // Make a real request to test access
+      // Make a minimal request just to check access
       const response = await axios.post(
         POST_IMAGE_GENERATIONS_URL,
         payload,
         {
           headers: OpenAIKeyChecker.getHeaders(key),
-          validateStatus: (status) => status >= 200 && status < 500, // Accept any non-server error response
-          timeout: 200000, // 200 second timeout to allow for image generation
-          signal: AbortSignal.timeout(200000)
+          validateStatus: (status) => true, // Accept any status code to inspect errors
+          timeout: 5000, // 5 second timeout
+          signal: AbortSignal.timeout(5000)
         }
       );
       
-      // Status 200 means success, but we need to verify the response actually contains image data
-      if (response.status === 200) {
-        // Check if the response contains the expected data structure
-        if (response.data && 
-            response.data.data && 
-            Array.isArray(response.data.data) && 
-            response.data.data.length > 0 && 
-            response.data.data[0].url) {
-          this.log.info(
-            { key: key.hash, status: response.status },
-            "Verified gpt-image-1 access: Successfully generated an image"
-          );
-          return true;
-        } else {
-          this.log.warn(
-            { key: key.hash, status: response.status, data: response.data },
-            "Response status is 200 but received unexpected response format. Image may not have been generated."
-          );
-          return false;
-        }
+      // If we get a 200, 400, or 500 response, consider the key valid
+      if (response.status === 200 || response.status === 400 || response.status === 500) {
+        this.log.info(
+          { key: key.hash, status: response.status },
+          `Verified gpt-image-1 access with status code ${response.status}`
+        );
+        return true;
       }
       
-      // All other status codes indicate issues
+      // Check for specific error responses that indicate no access
       const data = response.data as any;
+      const errorMessage = data?.error?.message || '';
+      
+      // Explicitly check for organization verification errors
+      if (response.status === 403 && errorMessage.includes("organization must be verified")) {
+        this.log.warn(
+          { key: key.hash, status: response.status, error: errorMessage },
+          "Key does not have access to gpt-image-1: organization verification required"
+        );
+        return false;
+      }
+      
+      // Only 401/403/404 responses indicate the key doesn't have model access
+      // (we're now considering 400 as valid)
+      if (response.status === 401 || response.status === 403 || response.status === 404) {
+        this.log.warn(
+          { key: key.hash, status: response.status, error: errorMessage },
+          `Key does not have access to gpt-image-1: received ${response.status} response`
+        );
+        return false;
+      }
+      
+      // For other status codes, log the issue but assume no access
       this.log.warn(
-        { key: key.hash, status: response.status, error: data?.error?.message || 'Unknown error' },
+        { key: key.hash, status: response.status, error: errorMessage },
         "Unexpected response when testing gpt-image-1 access, assuming no access"
       );
       return false;
       
     } catch (error) {
+      // Handle network errors or request failures
       if (error instanceof AxiosError && error.response) {
-        // Check for forbidden/unauthorized responses
         const status = error.response.status;
         const data = error.response.data as any;
         const errorMessage = data?.error?.message || 'Unknown error';
         
-        // Organization verification errors (403 forbidden)
-        if (status === 403) {
+        // Check for specific error messages related to access
+        if (errorMessage.includes("organization must be verified") || 
+            errorMessage.includes("does not have access") ||
+            errorMessage.includes("not available")) {
           this.log.warn(
-            { key: key.hash, error: errorMessage },
-            "Key does not have access to gpt-image-1: received 403 Forbidden response"
+            { key: key.hash, status, error: errorMessage },
+            "Key does not have access to gpt-image-1 based on error message"
           );
           return false;
         }
         
-        // Explicitly handle all status codes that indicate the key doesn't have access
-        if (status === 401 || status === 404 || status === 400) {
-          this.log.warn(
+        // Status 400 or 500 indicates the API processed the request but had validation issues
+        // This means the key is authorized to use the model
+        if (status === 400 || status === 500) {
+          this.log.info(
             { key: key.hash, status, error: errorMessage },
-            `Key does not have access to gpt-image-1: received ${status} response`
+            `Verified gpt-image-1 access with error response ${status}`
           );
-          return false;
+          return true;
         }
         
-        // Rate limit errors - we no longer assume these mean the key has access
-        // since the rate limit could be happening before model validation
-        if (status === 429) {
+        // 401/403/404 responses indicate the key doesn't have proper access
+        if (status === 401 || status === 403 || status === 404) {
           this.log.warn(
             { key: key.hash, status, error: errorMessage },
-            "Cannot verify gpt-image-1 access: key is rate limited. Assuming no access for safety."
+            `Key does not have access to gpt-image-1: received ${status} error response`
           );
           return false;
         }
@@ -433,7 +442,7 @@ export class OpenAIKeyChecker extends KeyCheckerBase<OpenAIKey> {
       // For all other errors, assume no access
       this.log.error(
         { key: key.hash, error: error.message },
-        "Unexpected error testing gpt-image-1 access. Assuming no access for safety."
+        "Error testing gpt-image-1 access. Assuming no access for safety."
       );
       return false;
     }