diff --git a/.changeset/chatty-pans-create.md b/.changeset/chatty-pans-create.md
new file mode 100644
index 000000000..45c08a23b
--- /dev/null
+++ b/.changeset/chatty-pans-create.md
@@ -0,0 +1,5 @@
+---
+"@browserbasehq/stagehand-evals": patch
+---
+
+Paramaterize execution model in agent evals
diff --git a/.env.example b/.env.example
index b31704e32..792df1fc5 100644
--- a/.env.example
+++ b/.env.example
@@ -11,3 +11,4 @@ EXPERIMENTAL_EVAL_MODELS="gpt-4o,claude-3-5-sonnet-latest,o1-mini,o1-preview"
 EVAL_CATEGORIES="observe,act,combination,extract,experimental"
 AGENT_EVAL_MAX_STEPS=50
 STAGEHAND_API_URL="http://localhost:80"
+AGENT_EVAL_EXECUTION_MODEL="google/gemini-2.5-flash
\ No newline at end of file
diff --git a/evals/initStagehand.ts b/evals/initStagehand.ts
index 519636fdc..f7b21d685 100644
--- a/evals/initStagehand.ts
+++ b/evals/initStagehand.ts
@@ -115,8 +115,10 @@ export const initStagehand = async ({
       provider: modelName.startsWith("claude") ? "anthropic" : "openai",
     } as AgentConfig;
   } else {
+    const executionModelFromEnv = process.env.AGENT_EVAL_EXECUTION_MODEL;
     agentConfig = {
       model: modelName,
+      executionModel: executionModelFromEnv ?? "google/gemini-2.5-flash",
     } as AgentConfig;
   }