Update the frameless realtime default model (#40321)

## What changed

- Use `gpt-live-1-codex` as the default model for frameless realtime sessions.
- Preserve an explicitly supplied per-session model override.

## Testing

- Cover both default and explicit model selection in the WebRTC frameless conversation test.
- Update the app-server live session request expectation for the new default.

GitOrigin-RevId: ce5dc978911ff0289d677b60fae2dd8cb1058d22

codex-rs/app-server/tests/suite/v2/realtime_conversation.rs

@@ -1907,7 +1907,7 @@ async fn webrtc_v3_start_posts_live_session_and_joins_without_session_update() -
     assert_call_create_multipart(
         harness.call_capture.single_request(),
         "v=offer\r\n",
-        r#"{"audio":{"output":{"voice":"cove"}},"delegation":{"ack_filler":false,"type":"client"},"instructions":"backend prompt\n\nstartup context","model":"gpt-live-1-boulder-alpha"}"#,
+        r#"{"audio":{"output":{"voice":"cove"}},"delegation":{"ack_filler":false,"type":"client"},"instructions":"backend prompt\n\nstartup context","model":"gpt-live-1-codex"}"#,
         "/v1/live",
     )?;
     assert!(

codex-rs/core/src/realtime_conversation.rs

@@ -102,7 +102,7 @@ const HANDOFF_STREAM_TRUNCATION_MARKER: &str = "\n…output truncated…\n";
 const AGENT_FINAL_MESSAGE_PREFIX: &str = "\"Agent Final Message\":\n\n";
 const STANDALONE_HANDOFF_ID: &str = "codex";
 const DEFAULT_REALTIME_MODEL: &str = "gpt-realtime-1.5";
-const DEFAULT_FRAMELESS_REALTIME_MODEL: &str = "gpt-live-1-boulder-alpha";
+const DEFAULT_FRAMELESS_REALTIME_MODEL: &str = "gpt-live-1-codex";
 pub(crate) const REALTIME_USER_TEXT_PREFIX: &str = "[USER] ";
 pub(crate) const REALTIME_BACKEND_TEXT_PREFIX: &str = "[BACKEND] ";
 const REALTIME_V2_HANDOFF_COMPLETE_ACKNOWLEDGEMENT: &str =

codex-rs/core/tests/suite/realtime_conversation.rs

@@ -588,8 +588,13 @@ async fn conversation_existing_call_attaches_without_creating_another_call() ->
     Ok(())
 }
 
+#[test_case(None, "gpt-live-1-codex"; "default model")]
+#[test_case(Some("session-override-model"), "session-override-model"; "explicit model")]
 #[tokio::test(flavor = "multi_thread", worker_threads = 2)]
-async fn conversation_webrtc_frameless_chatgpt_sends_codex_headers_to_backend() -> Result<()> {
+async fn conversation_webrtc_frameless_chatgpt_sends_codex_headers_to_backend(
+    model: Option<&str>,
+    expected_model: &str,
+) -> Result<()> {
     skip_if_no_network!(Ok(()));
 
     let server = start_mock_server().await;
@@ -619,7 +624,7 @@ async fn conversation_webrtc_frameless_chatgpt_sends_codex_headers_to_backend()
             config.experimental_realtime_ws_backend_prompt = Some("backend prompt".to_string());
             config.experimental_realtime_ws_base_url = Some(realtime_ws_base_url);
         });
-    let test = builder.build(&server).await?;
+    let test = builder.build_with_auto_env(&server).await?;
 
     test.codex
         .submit(Op::RealtimeConversationStart(ConversationStartParams {
@@ -631,7 +636,7 @@ async fn conversation_webrtc_frameless_chatgpt_sends_codex_headers_to_backend()
             codex_response_handoff_mode:
                 codex_protocol::protocol::CodexResponseHandoffMode::Thinking,
             codex_response_handoff_channel_prefixes: None,
-            model: Some("session-override-model".to_string()),
+            model: model.map(str::to_string),
             output_modality: RealtimeOutputModality::Audio,
             include_startup_context: false,
             initial_items: Vec::new(),
@@ -690,10 +695,12 @@ async fn conversation_webrtc_frameless_chatgpt_sends_codex_headers_to_backend()
         json!({
             "sdp": body["sdp"],
             "delegation": body["session"]["delegation"]["type"],
+            "model": body["session"]["model"],
         }),
         json!({
             "sdp": "v=offer\r\n",
             "delegation": "client",
+            "model": expected_model,
         })
     );