diff --git a/core/llm/fetchModels.ts b/core/llm/fetchModels.ts
index 88fe1946f95..efbff8bca97 100644
--- a/core/llm/fetchModels.ts
+++ b/core/llm/fetchModels.ts
@@ -179,8 +179,11 @@ async function fetchGeminiModels(
 ): Promise<FetchedModel[]> {
   const base = apiBase || "https://generativelanguage.googleapis.com/v1beta/";
   const url = new URL("models", base);
-  url.searchParams.set("key", apiKey ?? "");
-  const response = await fetch(url);
+  const response = await fetch(url, {
+    headers: {
+      "x-goog-api-key": apiKey ?? "",
+    },
+  });
   if (!response.ok) {
     throw new Error(`Failed to fetch Gemini models: ${response.status}`);
   }
diff --git a/extensions/cli/src/smoke-api/smoke-api-helpers.ts b/extensions/cli/src/smoke-api/smoke-api-helpers.ts
index ee2dcc922d7..2d00f50b96a 100644
--- a/extensions/cli/src/smoke-api/smoke-api-helpers.ts
+++ b/extensions/cli/src/smoke-api/smoke-api-helpers.ts
@@ -48,9 +48,9 @@ export async function cleanupSmokeContext(
   }
 }
 
-// Override via SMOKE_MODEL env var. Falls back to claude-3-haiku which is
-// widely available and cheap.
-const SMOKE_MODEL = process.env.SMOKE_MODEL || "claude-3-haiku-20240307";
+// Override via SMOKE_MODEL env var. Falls back to claude-haiku-4-5-20251001
+// which is widely available and cost-effective.
+const SMOKE_MODEL = process.env.SMOKE_MODEL || "claude-haiku-4-5-20251001";
 
 /**
  * Writes a YAML config that points at the real Anthropic API.