studio: fix GGUF download UX -- progress bar, cancel, sorting, auto-scroll

- Run GGUF load_model in asyncio.to_thread so the event loop stays free
  for progress polling during download (was blocking all requests).
- Extract download phase out of the lock in LlamaCppBackend.load_model
  so unload_model/cancel can take effect immediately during download.
- Fix "downloaded" badge for split GGUFs: check total cached bytes
  across all shards vs expected size, not just first shard existence.
- Respect CUDA_VISIBLE_DEVICES in /api/system GPU reporting so the
  frontend GGUF fit estimation uses actual available VRAM.
- Sort tight variants (need CPU offload) smallest-first instead of
  largest-first -- closer to GPU budget = faster inference.
- Fix cancel: use refs instead of React state for abort controller and
  toast ID so both cancel buttons (text + toast) work reliably. Make
  cancel synchronous (fire-and-forget unload) for instant UI response.
  Check abortCtrl.signal.aborted after loadModel returns to prevent
  ghost model state. Skip rollback and suppress errors on cancel.
- Dynamic top 4 GGUF models fetched from HF API sorted by downloads,
  prepended to the default recommended list.
- Remove turnAnchor="top" for auto-scroll to bottom during generation.
- Set default toast duration to 10s (was infinite for loading toasts).
- Deduplicate cached GGUF repos using scan_cache_dir API (fixes
  Qwen/X-GGUF vs qwen/x-gguf duplicates from lowercased HF cache).
- Pre-compile repo_id validation regex to silence CodeQL ReDoS warning.
- Change welcome text and default suggestion text.

studio/frontend/src/components/assistant-ui/model-selector/pickers.tsx

@@ -250,8 +250,10 @@ function GgufVariantExpander({
       const bIsRec = b.quant === effectiveRecommended;
       if (aIsRec !== bIsRec) return aIsRec ? -1 : 1;
 
-      // fits/tight: largest first (best quality); OOM: smallest first
-      return aTier === 4 ? a.size_bytes - b.size_bytes : b.size_bytes - a.size_bytes;
+      // fits: largest first (best quality that fits in GPU)
+      // tight/OOM: smallest first (closest to fitting, fastest to run)
+      const fitsInGpu = aTier === 0 || aTier === 2;
+      return fitsInGpu ? b.size_bytes - a.size_bytes : a.size_bytes - b.size_bytes;
     });
   }, [variants, effectiveRecommended, getGgufFit]);
 

studio/frontend/src/components/assistant-ui/thread.tsx

@@ -62,7 +62,6 @@ export const Thread: FC<{ hideComposer?: boolean; hideWelcome?: boolean }> = ({
       }}
     >
       <ThreadPrimitive.Viewport
-        turnAnchor="top"
         className="aui-thread-viewport relative flex flex-1 flex-col overflow-x-auto overflow-y-scroll scroll-smooth px-4 pt-4"
       >
         {!hideWelcome && (
@@ -140,10 +139,10 @@ const ThreadWelcome: FC<{ hideComposer?: boolean }> = ({ hideComposer }) => {
               className="size-20"
             />
             <h1 className="aui-thread-welcome-message-inner fade-in slide-in-from-bottom-1 animate-in font-semibold text-2xl duration-200">
-              Test Your Fine-tuned Model
+              Run LLMs or test your fine-tune
             </h1>
             <p className="aui-thread-welcome-message-inner fade-in slide-in-from-bottom-1 animate-in text-muted-foreground text-base delay-75 duration-200">
-              Start a conversation to see how your model performs.
+              Run GGUFs, safetensors, vision and audio models!
             </p>
           </div>
           <div className="grid grid-cols-2 gap-2">

studio/frontend/src/components/ui/sonner.tsx

@@ -1,6 +1,6 @@
-// SPDX-License-Identifier: AGPL-3.0-only
-// Copyright 2026-present the Unsloth AI Inc. team. All rights reserved. See /studio/LICENSE.AGPL-3.0
-
+// SPDX-License-Identifier: AGPL-3.0-only
+// Copyright 2026-present the Unsloth AI Inc. team. All rights reserved. See /studio/LICENSE.AGPL-3.0
+
 import {
   Alert02Icon,
   CheckmarkCircle02Icon,
@@ -19,6 +19,7 @@ const Toaster = ({ ...props }: ToasterProps) => {
     <Sonner
       theme={theme as ToasterProps["theme"]}
       className="toaster group"
+      duration={10000}
       icons={{
         success: (
           <HugeiconsIcon