Martin Bowling(@martinbowling):4B computer-use model. LoRA adapters on a frozen Qwen3.5-4B, SFT then RL where the only reward is finishing the task. 92.9% vs 60.1% on GUI-360, no accessibility tree. Small enough to run on-prem for real SMB workflows. Pulling it onto the Spark this week

Martin Bowling

@martinbowling

conquering the world, 280 characters at a time :) @GroqInc Groq Champ @Replit Builder in Residence Alumni.

加入 December 2007

2.1K 正在關注 4.7K 粉絲

Martin Bowling@martinbowling

2026.09.23 18:21

4B computer-use model. LoRA adapters on a frozen Qwen3.5-4B, SFT then RL where the only reward is finishing the task. 92.9% vs 60.1% on GUI-360, no accessibility tree. Small enough to run on-prem for real SMB workflows. Pulling it onto the Spark this week