fix: remove stream_options from xAI handler to fix Grok-4 API error

- Remove stream_options parameter from xAI streaming requests as the xAI API does not support it
- Add comment explaining why stream_options is not included
- Update tests to not expect stream_options in xAI API calls
- Add specific test case for Grok-4 streaming

Fixes #6702

src/api/providers/__tests__/xai.spec.ts

@@ -276,8 +276,35 @@ describe("XAIHandler", () => {
 				temperature: 0,
 				messages: expect.arrayContaining([{ role: "system", content: systemPrompt }]),
 				stream: true,
-				stream_options: { include_usage: true },
+				// xAI API doesn't support stream_options
 			}),
 		)
 	})
+
+	it("should not include stream_options for Grok-4 model", async () => {
+		// Create handler with Grok-4 model
+		const grok4Handler = new XAIHandler({ apiModelId: "grok-4" })
+
+		// Setup mock for streaming response
+		mockCreate.mockImplementationOnce(() => {
+			return {
+				[Symbol.asyncIterator]: () => ({
+					async next() {
+						return { done: true }
+					},
+				}),
+			}
+		})
+
+		// Start generating a message
+		const messageGenerator = grok4Handler.createMessage("test prompt", [])
+		await messageGenerator.next()
+
+		// Verify that stream_options was NOT included
+		const calls = mockCreate.mock.calls
+		const lastCall = calls[calls.length - 1][0]
+		expect(lastCall).not.toHaveProperty("stream_options")
+		expect(lastCall.stream).toBe(true)
+		expect(lastCall.model).toBe("grok-4")
+	})
 })

src/api/providers/xai.ts

@@ -54,7 +54,7 @@ export class XAIHandler extends BaseProvider implements SingleCompletionHandler
 			temperature: this.options.modelTemperature ?? XAI_DEFAULT_TEMPERATURE,
 			messages: [{ role: "system", content: systemPrompt }, ...convertToOpenAiMessages(messages)],
 			stream: true,
-			stream_options: { include_usage: true },
+			// xAI API doesn't support stream_options parameter
 			...(reasoning && reasoning),
 		})
 
@@ -75,15 +75,19 @@ export class XAIHandler extends BaseProvider implements SingleCompletionHandler
 				}
 			}
 
+			// Note: xAI API doesn't support stream_options, so usage data might not be available in streaming responses
 			if (chunk.usage) {
 				// Extract detailed token information if available
 				// First check for prompt_tokens_details structure (real API response)
-				const promptDetails = "prompt_tokens_details" in chunk.usage ? chunk.usage.prompt_tokens_details : null;
-				const cachedTokens = promptDetails && "cached_tokens" in promptDetails ? promptDetails.cached_tokens : 0;
+				const promptDetails = "prompt_tokens_details" in chunk.usage ? chunk.usage.prompt_tokens_details : null
+				const cachedTokens = promptDetails && "cached_tokens" in promptDetails ? promptDetails.cached_tokens : 0
 
 				// Fall back to direct fields in usage (used in test mocks)
-				const readTokens = cachedTokens || ("cache_read_input_tokens" in chunk.usage ? (chunk.usage as any).cache_read_input_tokens : 0);
-				const writeTokens = "cache_creation_input_tokens" in chunk.usage ? (chunk.usage as any).cache_creation_input_tokens : 0;
+				const readTokens =
+					cachedTokens ||
+					("cache_read_input_tokens" in chunk.usage ? (chunk.usage as any).cache_read_input_tokens : 0)
+				const writeTokens =
+					"cache_creation_input_tokens" in chunk.usage ? (chunk.usage as any).cache_creation_input_tokens : 0
 
 				yield {
 					type: "usage",