diff --git a/apps/studio/evals/assistant.eval.ts b/apps/studio/evals/assistant.eval.ts
index b21387a6bbf..a5fedcc9b67 100644
--- a/apps/studio/evals/assistant.eval.ts
+++ b/apps/studio/evals/assistant.eval.ts
@@ -1,51 +1,22 @@
 import { openai } from '@ai-sdk/openai'
 import { Eval } from 'braintrust'
-import { stripIndent } from 'common-tags'
 import {
-  AssistantEvalCase,
   criteriaMetScorer,
   sqlSimilarityScorer,
   textIncludesScorer,
   toolUsageScorer,
-} from 'lib/ai/evals/scorer'
+} from './scorer'
 import { generateAssistantResponse } from 'lib/ai/generate-assistant-response'
 import { getMockTools } from 'lib/ai/tools/mock-tools'
 import assert from 'node:assert'
+import { dataset } from './dataset'
 
 assert(process.env.BRAINTRUST_PROJECT_ID, 'BRAINTRUST_PROJECT_ID is not set')
 assert(process.env.OPENAI_API_KEY, 'OPENAI_API_KEY is not set')
 
 Eval('Assistant', {
   projectId: process.env.BRAINTRUST_PROJECT_ID,
-  data: () => {
-    return [
-      {
-        input: 'Hello!',
-        expected: { textIncludes: 'Hi' },
-        metadata: { category: ['other'] },
-      },
-      {
-        input: 'How do I implement IP address rate limiting?',
-        expected: { requiredTools: ['search_docs'] },
-        metadata: { category: ['general_help'] },
-      },
-      {
-        input: 'Check if my project is having issues right now and tell me what to fix first.',
-        expected: {
-          requiredTools: ['get_advisors', 'get_logs'],
-          criteria: 'Response reflects there are RLS issues to fix.',
-        },
-        metadata: { category: ['debugging', 'rls_policies'] },
-      },
-      {
-        input: 'Create a new table "foods" with columns for "name" and "color"',
-        expected: {
-          sqlQuery: stripIndent`CREATE TABLE IF NOT EXISTS public.foods ( id bigserial PRIMARY KEY, name text NOT NULL, color text );`,
-        },
-        metadata: { category: ['sql_generation', 'schema_design'] },
-      },
-    ] satisfies AssistantEvalCase[]
-  },
+  data: () => dataset,
   task: async (input) => {
     const result = await generateAssistantResponse({
       model: openai('gpt-5-mini'),
diff --git a/apps/studio/evals/dataset.ts b/apps/studio/evals/dataset.ts
new file mode 100644
index 00000000000..f7c5aa7a3bf
--- /dev/null
+++ b/apps/studio/evals/dataset.ts
@@ -0,0 +1,30 @@
+import { stripIndent } from 'common-tags'
+import { AssistantEvalCase } from './scorer'
+
+export const dataset: AssistantEvalCase[] = [
+  {
+    input: 'Hello!',
+    expected: { textIncludes: 'Hi' },
+    metadata: { category: ['other'] },
+  },
+  {
+    input: 'How do I implement IP address rate limiting?',
+    expected: { requiredTools: ['search_docs'] },
+    metadata: { category: ['general_help'] },
+  },
+  {
+    input: 'Check if my project is having issues right now and tell me what to fix first.',
+    expected: {
+      requiredTools: ['get_advisors', 'get_logs'],
+      criteria: 'Response reflects there are RLS issues to fix.',
+    },
+    metadata: { category: ['debugging', 'rls_policies'] },
+  },
+  {
+    input: 'Create a new table "foods" with columns for "name" and "color"',
+    expected: {
+      sqlQuery: stripIndent`CREATE TABLE IF NOT EXISTS public.foods ( id bigserial PRIMARY KEY, name text NOT NULL, color text );`,
+    },
+    metadata: { category: ['sql_generation', 'schema_design'] },
+  },
+]
diff --git a/apps/studio/lib/ai/evals/scorer.ts b/apps/studio/evals/scorer.ts
similarity index 99%
rename from apps/studio/lib/ai/evals/scorer.ts
rename to apps/studio/evals/scorer.ts
index ab7b7b073ab..a59b637861e 100644
--- a/apps/studio/lib/ai/evals/scorer.ts
+++ b/apps/studio/evals/scorer.ts
@@ -110,3 +110,4 @@ export const textIncludesScorer: EvalScorer<Input, Output, Expected> = async ({
     score: includes ? 1 : 0,
   }
 }
+