From 2ae30a4abb983645a2b91db3e88b5f605ca5733c Mon Sep 17 00:00:00 2001
From: John Onysko <johno10661@gmail.com>
Date: Sun, 24 May 2026 15:17:34 -0400
Subject: [PATCH] fix(chat): prefer selected model for suggestions, fall back
 to smallest
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

`getChatSuggestions` previously picked the largest installed model by file
size, on the assumption that bigger models give better suggestions. This
is unsafe: if any installed model exceeds available VRAM (e.g.
llama3.1:405b on a 96 GB GPU), Ollama spends minutes trying to load it
and the request 500s — making the chat page unusable for anyone who
happens to keep a flagship-sized model on disk.

Chat suggestions are short prompts that don't benefit from a flagship
model anyway. Prefer the user's selected `chat.lastModel` when set, and
fall back to the smallest installed model otherwise. `OllamaService.getModels()`
already excludes embedders, so the fallback always picks a chat model.
---
 admin/app/services/chat_service.ts | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

diff --git a/admin/app/services/chat_service.ts b/admin/app/services/chat_service.ts
index 2d97ad7..a5d07c9 100644
--- a/admin/app/services/chat_service.ts
+++ b/admin/app/services/chat_service.ts
@@ -1,5 +1,6 @@
 import ChatSession from '#models/chat_session'
 import ChatMessage from '#models/chat_message'
+import KVStore from '#models/kv_store'
 import logger from '@adonisjs/core/services/logger'
 import { DateTime } from 'luxon'
 import { inject } from '@adonisjs/core'
@@ -36,17 +37,23 @@ export class ChatService {
         return [] // If no models are available, return empty suggestions
       }
 
-      // Larger models generally give "better" responses, so pick the largest one
-      const largestModel = models.reduce((prev, current) => {
-        return prev.size > current.size ? prev : current
-      })
+      // Prefer the user's selected chat model. Fall back to the smallest
+      // installed model — picking the largest by file size is unsafe: if any
+      // installed model exceeds available VRAM (e.g. llama3.1:405b on a 96 GB
+      // GPU), Ollama spends minutes trying to load it and the request 500s.
+      // Suggestions are short prompts that don't benefit from a flagship model.
+      const lastModel = await KVStore.getValue('chat.lastModel')
+      const preferred = lastModel ? models.find((m) => m.name === lastModel) : undefined
+      const chosen =
+        preferred ??
+        models.reduce((prev, current) => (prev.size < current.size ? prev : current))
 
-      if (!largestModel) {
+      if (!chosen) {
         return []
       }
 
       const response = await this.ollamaService.chat({
-        model: largestModel.name,
+        model: chosen.name,
         messages: [
           {
             role: 'user',