diff --git a/dev-packages/node-integration-tests/suites/tracing/openai/orchestrion/test.ts b/dev-packages/node-integration-tests/suites/tracing/openai/orchestrion/test.ts
index 19bd448f4db0..2070b7c3f98b 100644
--- a/dev-packages/node-integration-tests/suites/tracing/openai/orchestrion/test.ts
+++ b/dev-packages/node-integration-tests/suites/tracing/openai/orchestrion/test.ts
@@ -37,7 +37,7 @@ describe('OpenAI integration (orchestrion)', () => {
             const chatSpans = container.items.filter(
               span => span.attributes[SEMANTIC_ATTRIBUTE_SENTRY_OP]?.value === 'gen_ai.chat',
             );
-            expect(chatSpans).toHaveLength(2);
+            expect(chatSpans).toHaveLength(3);
 
             const chatSpan = container.items.find(
               span => span.attributes[GEN_AI_RESPONSE_ID_ATTRIBUTE]?.value === 'chatcmpl-mock123',
@@ -92,6 +92,50 @@ describe('OpenAI integration (orchestrion)', () => {
               type: 'string',
               value: ORCHESTRION_ORIGIN,
             });
+
+            const responsesSpan = container.items.find(
+              span => span.attributes[GEN_AI_RESPONSE_ID_ATTRIBUTE]?.value === 'resp_mock456',
+            );
+            expect(responsesSpan).toBeDefined();
+            expect(responsesSpan!.name).toBe('chat gpt-3.5-turbo');
+            expect(responsesSpan!.status).toBe('ok');
+            expect(responsesSpan!.attributes[SEMANTIC_ATTRIBUTE_SENTRY_ORIGIN]).toEqual({
+              type: 'string',
+              value: ORCHESTRION_ORIGIN,
+            });
+            expect(responsesSpan!.attributes[SEMANTIC_ATTRIBUTE_SENTRY_OP]).toEqual({
+              type: 'string',
+              value: 'gen_ai.chat',
+            });
+            expect(responsesSpan!.attributes[GEN_AI_OPERATION_NAME_ATTRIBUTE]).toEqual({
+              type: 'string',
+              value: 'chat',
+            });
+            expect(responsesSpan!.attributes[GEN_AI_SYSTEM_ATTRIBUTE]).toEqual({ type: 'string', value: 'openai' });
+            expect(responsesSpan!.attributes[GEN_AI_REQUEST_MODEL_ATTRIBUTE]).toEqual({
+              type: 'string',
+              value: 'gpt-3.5-turbo',
+            });
+            expect(responsesSpan!.attributes[GEN_AI_RESPONSE_MODEL_ATTRIBUTE]).toEqual({
+              type: 'string',
+              value: 'gpt-3.5-turbo',
+            });
+            expect(responsesSpan!.attributes[GEN_AI_RESPONSE_FINISH_REASONS_ATTRIBUTE]).toEqual({
+              type: 'string',
+              value: '["completed"]',
+            });
+            expect(responsesSpan!.attributes[GEN_AI_USAGE_INPUT_TOKENS_ATTRIBUTE]).toEqual({
+              type: 'integer',
+              value: 5,
+            });
+            expect(responsesSpan!.attributes[GEN_AI_USAGE_OUTPUT_TOKENS_ATTRIBUTE]).toEqual({
+              type: 'integer',
+              value: 8,
+            });
+            expect(responsesSpan!.attributes[GEN_AI_USAGE_TOTAL_TOKENS_ATTRIBUTE]).toEqual({
+              type: 'integer',
+              value: 13,
+            });
           },
         })
         .start()
@@ -135,6 +179,23 @@ describe('OpenAI integration (orchestrion)', () => {
                 type: 'string',
                 value: '["Hello from OpenAI mock!"]',
               });
+
+              const responsesSpan = container.items.find(
+                span => span.attributes[GEN_AI_RESPONSE_ID_ATTRIBUTE]?.value === 'resp_mock456',
+              );
+              expect(responsesSpan).toBeDefined();
+              expect(responsesSpan!.attributes[GEN_AI_INPUT_MESSAGES_ORIGINAL_LENGTH_ATTRIBUTE]).toEqual({
+                type: 'integer',
+                value: 1,
+              });
+              expect(responsesSpan!.attributes[GEN_AI_INPUT_MESSAGES_ATTRIBUTE]).toEqual({
+                type: 'string',
+                value: 'Translate this to French: Hello',
+              });
+              expect(responsesSpan!.attributes[GEN_AI_RESPONSE_TEXT_ATTRIBUTE]).toEqual({
+                type: 'string',
+                value: 'Response to: Translate this to French: Hello',
+              });
             },
           })
           .start()
diff --git a/packages/server-utils/src/integrations/tracing-channel/openai.ts b/packages/server-utils/src/integrations/tracing-channel/openai.ts
index e024a380947c..1f5430bf11e7 100644
--- a/packages/server-utils/src/integrations/tracing-channel/openai.ts
+++ b/packages/server-utils/src/integrations/tracing-channel/openai.ts
@@ -24,7 +24,11 @@ const INTEGRATION_NAME = 'OpenAI' as const;
 // are attributable separately from the OTel/proxy one.
 const ORIGIN = 'auto.ai.orchestrion.openai';
 
-const OPERATION = 'chat';
+// Each instrumented `create` method maps to the gen_ai operation its span reports.
+const INSTRUMENTED_CHANNELS = [
+  { channel: CHANNELS.OPENAI_CHAT, operation: 'chat' },
+  { channel: CHANNELS.OPENAI_RESPONSES, operation: 'chat' },
+] as const;
 
 /**
  * The context object orchestrion shares across the tracing-channel lifecycle hooks: `arguments` is the
@@ -48,37 +52,38 @@ const _openaiChannelIntegration = ((options: OpenAiOptions = {}) => {
       }
       subscribed = true;
 
-      DEBUG_BUILD && debug.log(`[orchestrion:openai] subscribing to channel "${CHANNELS.OPENAI_CHAT}"`);
-
       // `bindTracingChannelToSpan` needs the async-context binding that `initOpenTelemetry()` registers
       // after `setupOnce` runs, so wait for it before subscribing.
       waitForTracingChannelBinding(() => {
-        bindTracingChannelToSpan(
-          diagnosticsChannel.tracingChannel<OpenAiChatChannelContext>(CHANNELS.OPENAI_CHAT),
-          data => createChatSpan(data, options),
-          {
-            beforeSpanEnd: (span, data) => {
-              if ('result' in data) {
-                addResponseAttributes(span, data.result, resolveAIRecordingOptions(options).recordOutputs);
-              }
+        for (const { channel, operation } of INSTRUMENTED_CHANNELS) {
+          DEBUG_BUILD && debug.log(`[orchestrion:openai] subscribing to channel "${channel}"`);
+          bindTracingChannelToSpan(
+            diagnosticsChannel.tracingChannel<OpenAiChatChannelContext>(channel),
+            data => createGenAiSpan(data, operation, options),
+            {
+              beforeSpanEnd: (span, data) => {
+                if ('result' in data) {
+                  addResponseAttributes(span, data.result, resolveAIRecordingOptions(options).recordOutputs);
+                }
+              },
+              captureError: () => ({ mechanism: { type: ORIGIN, handled: false } }),
             },
-            captureError: () => ({ mechanism: { type: ORIGIN, handled: false } }),
-          },
-        );
+          );
+        }
       });
     },
   };
 }) satisfies IntegrationFn;
 
 /**
- * Build the span for a `chat.completions.create` call.
+ * Build the span for an instrumented `create` call.
  * Returning `undefined` opts the payload out so no span is opened.
  */
-function createChatSpan(data: OpenAiChatChannelContext, options: OpenAiOptions): Span | undefined {
+function createGenAiSpan(data: OpenAiChatChannelContext, operation: string, options: OpenAiOptions): Span | undefined {
   const args = data.arguments ?? [];
   const params = args[0] as Record<string, unknown> | undefined;
 
-  // streaming is not supported yet
+  // streaming is not supported
   if (params?.stream === true) {
     return undefined;
   }
@@ -86,27 +91,26 @@ function createChatSpan(data: OpenAiChatChannelContext, options: OpenAiOptions):
   const { recordInputs } = resolveAIRecordingOptions(options);
   const enableTruncation = shouldEnableTruncation(options.enableTruncation);
 
-  const attributes = extractRequestAttributes(args, OPERATION);
+  const attributes = extractRequestAttributes(args, operation);
   attributes[SEMANTIC_ATTRIBUTE_SENTRY_ORIGIN] = ORIGIN;
   const model = (params?.model as string) || 'unknown';
 
   const span = startInactiveSpan({
-    name: `${OPERATION} ${model}`,
-    op: `gen_ai.${OPERATION}`,
+    name: `${operation} ${model}`,
+    op: `gen_ai.${operation}`,
     attributes: attributes as Record<string, SpanAttributeValue>,
   });
 
   if (recordInputs && params) {
-    addRequestAttributes(span, params, OPERATION, enableTruncation);
+    addRequestAttributes(span, params, operation, enableTruncation);
   }
 
   return span;
 }
 
 /**
- * EXPERIMENTAL — orchestrion-driven OpenAI integration. Subscribes to the `orchestrion:openai:chat`
- * diagnostics_channel injected into `openai`'s `Completions.prototype.create`, so it requires the
- * orchestrion runtime hook or bundler plugin. Covers non-streaming `chat.completions.create`; streaming
- * and the other methods are follow-ups. Browser/edge keep using the proxy `instrumentOpenAiClient`.
+ * EXPERIMENTAL — orchestrion-driven OpenAI integration. Subscribes to the `orchestrion:openai:*`
+ * diagnostics_channels injected into `openai`'s `create` methods, so it requires the orchestrion runtime
+ * hook or bundler plugin. Covers non-streaming `chat.completions.create` and `responses.create`.
  */
 export const openaiChannelIntegration = defineIntegration(_openaiChannelIntegration);
diff --git a/packages/server-utils/src/orchestrion/channels.ts b/packages/server-utils/src/orchestrion/channels.ts
index afba59a4c6bc..785f8aef8c53 100644
--- a/packages/server-utils/src/orchestrion/channels.ts
+++ b/packages/server-utils/src/orchestrion/channels.ts
@@ -15,6 +15,7 @@ export const CHANNELS = {
   MYSQL_QUERY: 'orchestrion:mysql:query',
   LRU_MEMOIZER_LOAD: 'orchestrion:lru-memoizer:load',
   OPENAI_CHAT: 'orchestrion:openai:chat',
+  OPENAI_RESPONSES: 'orchestrion:openai:responses',
 } as const;
 
 export type ChannelName = (typeof CHANNELS)[keyof typeof CHANNELS];
diff --git a/packages/server-utils/src/orchestrion/config.ts b/packages/server-utils/src/orchestrion/config.ts
index 96b2d2998df2..f1b41f802c34 100644
--- a/packages/server-utils/src/orchestrion/config.ts
+++ b/packages/server-utils/src/orchestrion/config.ts
@@ -46,6 +46,12 @@ export const SENTRY_INSTRUMENTATIONS: InstrumentationConfig[] = [
     module: { name: 'openai', versionRange: '>=4.0.0 <7', filePath },
     functionQuery: { className: 'Completions', methodName: 'create', kind: 'Auto' as const },
   })) satisfies InstrumentationConfig[]),
+  // OpenAI responses API — same `create(body, options)` shape as chat completions.
+  ...(['resources/responses/responses.js', 'resources/responses/responses.mjs'].map(filePath => ({
+    channelName: 'responses',
+    module: { name: 'openai', versionRange: '>=4.0.0 <7', filePath },
+    functionQuery: { className: 'Responses', methodName: 'create', kind: 'Auto' as const },
+  })) satisfies InstrumentationConfig[]),
 ];
 
 /**