freebuff: Enable kimi 24/7

Author: jahooma

web/src/llm-api/__tests__/fireworks-deployment.test.ts

@@ -184,10 +184,16 @@ describe('Fireworks deployment routing', () => {
       expect(fetchCalls).toEqual([KIMI_STANDARD_MODEL_ID])
     })
 
-    it('keeps Kimi unavailable outside hours when no deployment is mapped', async () => {
-      const mockFetch = mock(async () => {
-        throw new Error('should not fetch outside deployment hours')
-      }) as unknown as typeof globalThis.fetch
+    it('uses serverless API for Kimi outside deployment hours (Kimi is 24/7)', async () => {
+      const fetchCalls: string[] = []
+
+      const mockFetch = mock(
+        async (_url: string | URL | Request, init?: RequestInit) => {
+          const body = JSON.parse(init?.body as string)
+          fetchCalls.push(body.model)
+          return new Response(JSON.stringify({ ok: true }), { status: 200 })
+        },
+      ) as unknown as typeof globalThis.fetch
 
       const response = await createFireworksRequestWithFallback({
         body: kimiBody as never,
@@ -202,9 +208,8 @@ describe('Fireworks deployment routing', () => {
         now: BEFORE_DEPLOYMENT_HOURS,
       })
 
-      expect(response.status).toBe(503)
-      const body = await response.json()
-      expect(body.error.code).toBe('DEPLOYMENT_OUTSIDE_HOURS')
+      expect(response.status).toBe(200)
+      expect(fetchCalls).toEqual([KIMI_STANDARD_MODEL_ID])
     })
 
     it('keeps GLM unavailable outside hours when no deployment is mapped', async () => {
@@ -433,7 +438,7 @@ describe('Fireworks deployment routing', () => {
       expect(body.error.code).toBe('DEPLOYMENT_OUTSIDE_HOURS')
     })
 
-    it('falls back to the standard Fireworks API for Kimi lite mode outside deployment hours', async () => {
+    it('uses the standard Fireworks API for Kimi lite mode outside deployment hours', async () => {
       const fetchCalls: string[] = []
 
       const mockFetch = mock(

web/src/llm-api/fireworks-config.ts

@@ -10,8 +10,8 @@ export const FIREWORKS_ACCOUNT_ID = 'james-65d217'
 
 export const FIREWORKS_DEPLOYMENT_MAP: Record<string, string> = {
   // 'minimax/minimax-m2.5': 'accounts/james-65d217/deployments/lnfid5h9',
-  // Disabled: route Kimi K2.6 through the Fireworks serverless API during
-  // availability hours instead of the dedicated deployment.
+  // Disabled: route Kimi K2.6 through the Fireworks serverless API (24/7)
+  // instead of the dedicated deployment.
   // 'moonshotai/kimi-k2.6': 'accounts/james-65d217/deployments/mjb4i7ea',
   // 'minimax/minimax-m2.7': 'accounts/james-65d217/deployments/nrdudqxd',
 }

web/src/llm-api/fireworks.ts

@@ -40,10 +40,7 @@ const FIREWORKS_MODEL_MAP: Record<string, string> = {
 }
 
 /** Models that stay limited to freebuff deployment hours even on serverless. */
-const FIREWORKS_HOURS_GATED_MODELS = new Set<string>([
-  'moonshotai/kimi-k2.6',
-  'z-ai/glm-5.1',
-])
+const FIREWORKS_HOURS_GATED_MODELS = new Set<string>(['z-ai/glm-5.1'])
 
 /** Flag to enable custom Fireworks deployments (set to false to use global API only) */
 const FIREWORKS_USE_CUSTOM_DEPLOYMENT = true