Fix CAPI reasoning E2E fixtures

dotnet/test/E2E/SessionE2ETests.cs

@@ -615,10 +615,10 @@ public async Task Should_Set_Model_With_ReasoningEffort()
 
         var modelChangedTask = TestHelper.GetNextEventOfTypeAsync<SessionModelChangeEvent>(session);
 
-        await session.SetModelAsync("gpt-4.1", "high");
+        await session.SetModelAsync("gpt-5.4", "high");
 
         var modelChanged = await modelChangedTask;
-        Assert.Equal("gpt-4.1", modelChanged.Data.NewModel);
+        Assert.Equal("gpt-5.4", modelChanged.Data.NewModel);
         Assert.Equal("high", modelChanged.Data.ReasoningEffort);
     }
 

dotnet/test/E2E/StreamingFidelityE2ETests.cs

@@ -152,6 +152,7 @@ public async Task Should_Emit_Streaming_Deltas_With_Reasoning_Effort_Configured(
         // the streaming pipeline — deltas still arrive and complete successfully.
         var session = await CreateSessionAsync(new SessionConfig
         {
+            Model = "gpt-5.4",
             Streaming = true,
             ReasoningEffort = "high",
         });

go/internal/e2e/session_e2e_test.go

@@ -1072,15 +1072,15 @@ func TestSetModelWithReasoningEffortE2E(t *testing.T) {
 		}
 	})
 
-	if err := session.SetModel(t.Context(), "gpt-4.1", &copilot.SetModelOptions{ReasoningEffort: copilot.String("high")}); err != nil {
+	if err := session.SetModel(t.Context(), "gpt-5.4", &copilot.SetModelOptions{ReasoningEffort: copilot.String("high")}); err != nil {
 		t.Fatalf("SetModel returned error: %v", err)
 	}
 
 	select {
 	case evt := <-modelChanged:
 		md, mdOk := evt.Data.(*copilot.SessionModelChangeData)
-		if !mdOk || md.NewModel != "gpt-4.1" {
-			t.Errorf("Expected newModel 'gpt-4.1', got %v", evt.Data)
+		if !mdOk || md.NewModel != "gpt-5.4" {
+			t.Errorf("Expected newModel 'gpt-5.4', got %v", evt.Data)
 		}
 		if !mdOk || md.ReasoningEffort == nil || *md.ReasoningEffort != "high" {
 			t.Errorf("Expected reasoningEffort 'high', got %v", evt.Data)

go/internal/e2e/streaming_fidelity_e2e_test.go

@@ -291,6 +291,7 @@ func TestStreamingFidelityE2E(t *testing.T) {
 		// the streaming pipeline — deltas still arrive and complete successfully.
 		session, err := client.CreateSession(t.Context(), &copilot.SessionConfig{
 			OnPermissionRequest: copilot.PermissionHandler.ApproveAll,
+			Model:               "gpt-5.4",
 			Streaming:           copilot.Bool(true),
 			ReasoningEffort:     "high",
 		})

java/src/test/java/com/github/copilot/StreamingFidelityTest.java

@@ -254,7 +254,7 @@ void testShouldEmitStreamingDeltasWithReasoningEffortConfigured() throws Excepti
         try (CopilotClient client = ctx.createClient()) {
             CopilotSession session = client
                     .createSession(new SessionConfig().setOnPermissionRequest(PermissionHandler.APPROVE_ALL)
-                            .setStreaming(true).setReasoningEffort("high"))
+                            .setModel("gpt-5.4").setStreaming(true).setReasoningEffort("high"))
                     .get();
 
             List<SessionEvent> events = new ArrayList<>();

nodejs/test/e2e/session.e2e.test.ts

@@ -969,10 +969,10 @@ describe("Send Blocking Behavior", async () => {
 
         const modelChangePromise = getNextEventOfType(session, "session.model_change");
 
-        await session.setModel("gpt-4.1", { reasoningEffort: "high" });
+        await session.setModel("gpt-5.4", { reasoningEffort: "high" });
 
         const event = await modelChangePromise;
-        expect(event.data.newModel).toBe("gpt-4.1");
+        expect(event.data.newModel).toBe("gpt-5.4");
         expect(event.data.reasoningEffort).toBe("high");
     });
 });

nodejs/test/e2e/streaming_fidelity.e2e.test.ts

@@ -146,8 +146,11 @@ describe("Streaming Fidelity", async () => {
     });
 
     it("should emit streaming deltas with reasoning effort configured", async () => {
-        const session = await client.createSession({
+        const reasoningClient = createClient();
+        onTestFinished(() => reasoningClient.stop());
+        const session = await reasoningClient.createSession({
             onPermissionRequest: approveAll,
+            model: "gpt-5.4",
             streaming: true,
             reasoningEffort: "high",
         });

python/e2e/test_session_e2e.py

@@ -695,10 +695,10 @@ def on_event(event):
 
         session.on(on_event)
 
-        await session.set_model("gpt-4.1", reasoning_effort="high")
+        await session.set_model("gpt-5.4", reasoning_effort="high")
 
         data = await asyncio.wait_for(model_change_event, timeout=30)
-        assert data.new_model == "gpt-4.1"
+        assert data.new_model == "gpt-5.4"
         assert data.reasoning_effort == "high"
 
     async def test_should_accept_blob_attachments(self, ctx: E2ETestContext):

python/e2e/test_streaming_fidelity_e2e.py

@@ -163,6 +163,7 @@ async def test_should_emit_streaming_deltas_with_reasoning_effort_configured(
 
         session = await ctx.client.create_session(
             on_permission_request=PermissionHandler.approve_all,
+            model="gpt-5.4",
             streaming=True,
             reasoning_effort="high",
         )

rust/tests/e2e/session.rs

@@ -967,7 +967,7 @@ async fn should_set_model_with_reasoningeffort() {
 
             session
                 .set_model(
-                    "gpt-4.1",
+                    "gpt-5.4",
                     Some(SetModelOptions::default().with_reasoning_effort("high")),
                 )
                 .await
@@ -976,7 +976,7 @@ async fn should_set_model_with_reasoningeffort() {
             let data = event
                 .typed_data::<SessionModelChangeData>()
                 .expect("session.model_change data");
-            assert_eq!(data.new_model, "gpt-4.1");
+            assert_eq!(data.new_model, "gpt-5.4");
             assert_eq!(data.reasoning_effort.as_deref(), Some("high"));
 
             session.disconnect().await.expect("disconnect session");

rust/tests/e2e/streaming_fidelity.rs

@@ -237,6 +237,7 @@ async fn should_emit_streaming_deltas_with_reasoning_effort_configured() {
                 let session = client
                     .create_session(
                         ctx.approve_all_session_config()
+                            .with_model("gpt-5.4")
                             .with_streaming(true)
                             .with_reasoning_effort("high"),
                     )

test/snapshots/streaming_fidelity/should_emit_streaming_deltas_with_reasoning_effort_configured.yaml

@@ -1,5 +1,5 @@
 models:
-  - claude-sonnet-4.5
+  - gpt-5.4
 conversations:
   - messages:
       - role: system