Understand this story
Meta and Carnegie Mellon University published a paper finding that small models trained with reinforcement learning select the correct tool more often than the prompt-driven 402B Llama-4-Maverick. The researchers trained Qwen3 models to break requests into steps, search step by step with keywords, and select 1 tool per step, first training on teacher-written examples, then using reinforcement learning to reward effective search steps and precise selection. On ToolBench, the trained 8B model selected the correct tool precisely 41.5% of the time, versus 30.6% for Maverick's best prompt setting; without retraining, it surpassed Maverick on 3 of 4 new tool benchmarks, and further improved after switching to a stronger retriever.
AI-generated from public reports
Developments
1 confirmed developments
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Oct 24
Rohan Paul
Meta and Carnegie Mellon University paper: RL-trained small model surpasses 402B Llama-4-Maverick in tool selection accuracy
A new paper from Meta and Carnegie Mellon University finds that a small model trained with reinforcement learning selects the correct tool more often than the prompt-driven 402B Llama-4-Maverick. The researchers trained a Qwen3 model to break requests into steps, search step by step with keywords, and select 1 tool per step; training first used teacher-written examples, then reinforcement learning rewarded effective search steps and precise selection. On ToolBench, the trained 8B model selected the correct tool precisely 41.5% of the time, versus 30.6% for Maverick's best prompt setting; without retraining, it surpassed Maverick on 3 of 4 new tool benchmarks, and further improved after switching to a stronger retriever.