updates to prompt caching

Author: Sg312

apps/sim/lib/copilot/chat/post.test.ts

@@ -167,8 +167,15 @@ describe('handleUnifiedChatPost', () => {
     )
 
     expect(response.status).toBe(200)
+    expect(buildCopilotRequestPayload).toHaveBeenCalledWith(
+      expect.objectContaining({
+        model: 'claude-opus-4-7',
+      }),
+      { selectedModel: 'claude-opus-4-7' }
+    )
     expect(createSSEStream).toHaveBeenCalledWith(
       expect.objectContaining({
+        titleModel: 'claude-opus-4-7',
         workspaceId: 'ws-1',
         orchestrateOptions: expect.objectContaining({
           workflowId: 'wf-1',
@@ -206,6 +213,7 @@ describe('handleUnifiedChatPost', () => {
     )
     expect(createSSEStream).toHaveBeenCalledWith(
       expect.objectContaining({
+        titleModel: 'claude-opus-4-7',
         workspaceId: 'ws-1',
         orchestrateOptions: expect.objectContaining({
           workspaceId: 'ws-1',

apps/sim/lib/copilot/chat/post.ts

@@ -50,7 +50,7 @@ import type { ChatContext } from '@/stores/panel'
 export const maxDuration = 3600
 
 const logger = createLogger('UnifiedChatAPI')
-const DEFAULT_MODEL = 'claude-opus-4-6'
+const DEFAULT_MODEL = 'claude-opus-4-7'
 
 const FileAttachmentSchema = z.object({
   id: z.string(),
@@ -147,6 +147,7 @@ type UnifiedChatBranch =
       workflowId: string
       workflowName?: string
       workspaceId?: string
+      effectiveModel: string
       selectedModel: string
       mode: UnifiedChatRequest['mode']
       provider?: string
@@ -182,6 +183,7 @@ type UnifiedChatBranch =
   | {
       kind: 'workspace'
       workspaceId: string
+      effectiveModel: string
       goRoute: '/api/mothership'
       titleModel: string
       titleProvider?: undefined
@@ -554,6 +556,7 @@ async function resolveBranch(params: {
       workflowId: resolvedWorkflowId,
       workflowName: resolved.workflowName,
       workspaceId: resolvedWorkspaceId,
+        effectiveModel: selectedModel,
       selectedModel,
       mode: mode ?? 'agent',
       provider,
@@ -620,6 +623,7 @@ async function resolveBranch(params: {
   return {
     kind: 'workspace',
     workspaceId: requestedWorkspaceId,
+    effectiveModel: DEFAULT_MODEL,
     goRoute: '/api/mothership',
     titleModel: DEFAULT_MODEL,
     notifyWorkspaceStatus: true,
@@ -821,7 +825,7 @@ export async function handleUnifiedChatPost(req: NextRequest) {
       activeOtelRoot.setRequestShape({
         branchKind: branch.kind,
         mode: body.mode,
-        model: body.model,
+        model: branch.effectiveModel,
         provider: body.provider,
         createNewChat: body.createNewChat,
         prefetch: body.prefetch,

apps/sim/lib/copilot/generated/trace-attributes-v1.ts

@@ -325,23 +325,38 @@ export const TraceAttr = {
   GenAiRequestMessagesCount: 'gen_ai.request.messages.count',
   GenAiRequestModel: 'gen_ai.request.model',
   GenAiRequestPromptCacheBreakpointCreated: 'gen_ai.request.prompt_cache.breakpoint.created',
+  GenAiRequestPromptCacheBreakpointDynamicWriteSuppressReason:
+    'gen_ai.request.prompt_cache.breakpoint.dynamic_write_suppress_reason',
+  GenAiRequestPromptCacheBreakpointDynamicWriteSuppressed:
+    'gen_ai.request.prompt_cache.breakpoint.dynamic_write_suppressed',
   GenAiRequestPromptCacheBreakpointEligible: 'gen_ai.request.prompt_cache.breakpoint.eligible',
   GenAiRequestPromptCacheBreakpointKind: 'gen_ai.request.prompt_cache.breakpoint.kind',
+  GenAiRequestPromptCacheBreakpointMinimumTokens:
+    'gen_ai.request.prompt_cache.breakpoint.minimum_tokens',
   GenAiRequestPromptCacheBreakpointPrefixTokensEstimated:
     'gen_ai.request.prompt_cache.breakpoint.prefix_tokens_estimated',
+  GenAiRequestPromptCacheBreakpointPreviousBoundaryTokens:
+    'gen_ai.request.prompt_cache.breakpoint.previous_boundary_tokens',
   GenAiRequestPromptCacheBreakpointSection: 'gen_ai.request.prompt_cache.breakpoint.section',
   GenAiRequestPromptCacheBreakpointSkipReason: 'gen_ai.request.prompt_cache.breakpoint.skip_reason',
   GenAiRequestPromptCacheBreakpointTargetIndex:
     'gen_ai.request.prompt_cache.breakpoint.target_index',
   GenAiRequestPromptCacheBreakpointTargetRole: 'gen_ai.request.prompt_cache.breakpoint.target_role',
   GenAiRequestPromptCacheBreakpointTargetType: 'gen_ai.request.prompt_cache.breakpoint.target_type',
+  GenAiRequestPromptCacheBreakpointThresholdTokens:
+    'gen_ai.request.prompt_cache.breakpoint.threshold_tokens',
   GenAiRequestPromptCacheBreakpointTtl: 'gen_ai.request.prompt_cache.breakpoint.ttl',
   GenAiRequestPromptCacheBreakpointsCount: 'gen_ai.request.prompt_cache.breakpoints.count',
   GenAiRequestPromptCacheBreakpointsCreated: 'gen_ai.request.prompt_cache.breakpoints.created',
+  GenAiRequestPromptCacheDynamicMinimumTokens: 'gen_ai.request.prompt_cache.dynamic.minimum_tokens',
   GenAiRequestPromptCacheDynamicPrefixTokensEstimated:
     'gen_ai.request.prompt_cache.dynamic.prefix_tokens_estimated',
+  GenAiRequestPromptCacheDynamicPreviousBoundaryTokens:
+    'gen_ai.request.prompt_cache.dynamic.previous_boundary_tokens',
   GenAiRequestPromptCacheDynamicSkipReason: 'gen_ai.request.prompt_cache.dynamic.skip_reason',
   GenAiRequestPromptCacheDynamicTargetType: 'gen_ai.request.prompt_cache.dynamic.target_type',
+  GenAiRequestPromptCacheDynamicThresholdTokens:
+    'gen_ai.request.prompt_cache.dynamic.threshold_tokens',
   GenAiRequestPromptCacheDynamicTtl: 'gen_ai.request.prompt_cache.dynamic.ttl',
   GenAiRequestPromptCacheDynamicUserEligible: 'gen_ai.request.prompt_cache.dynamic_user_eligible',
   GenAiRequestPromptCacheDynamicUserSkipReason:
@@ -903,20 +918,28 @@ export const TraceAttrValues: readonly TraceAttrValue[] = [
   'gen_ai.request.messages.count',
   'gen_ai.request.model',
   'gen_ai.request.prompt_cache.breakpoint.created',
+  'gen_ai.request.prompt_cache.breakpoint.dynamic_write_suppress_reason',
+  'gen_ai.request.prompt_cache.breakpoint.dynamic_write_suppressed',
   'gen_ai.request.prompt_cache.breakpoint.eligible',
   'gen_ai.request.prompt_cache.breakpoint.kind',
+  'gen_ai.request.prompt_cache.breakpoint.minimum_tokens',
   'gen_ai.request.prompt_cache.breakpoint.prefix_tokens_estimated',
+  'gen_ai.request.prompt_cache.breakpoint.previous_boundary_tokens',
   'gen_ai.request.prompt_cache.breakpoint.section',
   'gen_ai.request.prompt_cache.breakpoint.skip_reason',
   'gen_ai.request.prompt_cache.breakpoint.target_index',
   'gen_ai.request.prompt_cache.breakpoint.target_role',
   'gen_ai.request.prompt_cache.breakpoint.target_type',
+  'gen_ai.request.prompt_cache.breakpoint.threshold_tokens',
   'gen_ai.request.prompt_cache.breakpoint.ttl',
   'gen_ai.request.prompt_cache.breakpoints.count',
   'gen_ai.request.prompt_cache.breakpoints.created',
+  'gen_ai.request.prompt_cache.dynamic.minimum_tokens',
   'gen_ai.request.prompt_cache.dynamic.prefix_tokens_estimated',
+  'gen_ai.request.prompt_cache.dynamic.previous_boundary_tokens',
   'gen_ai.request.prompt_cache.dynamic.skip_reason',
   'gen_ai.request.prompt_cache.dynamic.target_type',
+  'gen_ai.request.prompt_cache.dynamic.threshold_tokens',
   'gen_ai.request.prompt_cache.dynamic.ttl',
   'gen_ai.request.prompt_cache.dynamic_user_eligible',
   'gen_ai.request.prompt_cache.dynamic_user_skip_reason',