Russet

Please ensure you have updated Russet to its latest release:

Milestones

• Export and import chats with local models across devices with Russet v2026.05.02+

• Alibaba's Qwen3.6-35B-A3B and Qwen3.6-27B models now available in v2026.04.30 for Mac. Next-level for coding and other demanding tasks. Especially useful to pair with Russet's Local Inference Server feature (Settings > Developer).

• PrismML's ultra dense intelligent Ternary Bonsai models are now available in v2026.04.27. They're super efficient yet smart which makes them ideal especially for iPhone and iPad.

Bugs

• 2026.04.11, 1 AM GMT: I am aware that smaller Gemma 4 models may generate some gibberish in their response. On Mac, larger Qwen3.5 models will also get improved performance with distinct reasoning traces. Both fixes will appear soon in v2026.04.24. ✅ Resolved

Tips

• Try the new Zen vibe for a simple monochrome UI. Pairs excellent with lowered content opacity on Russet for Mac.