Multi-Harness RL Boosts LFM2.5 from 42% to 54%

The same LFM2.5 model performs nearly twice as differently across agent harnesses; a multi-harness RL guide from Hugging Face and Adithya S K shows RL training lifts its score from 42% to 54% while cutting tool calls by 31%.

2026-10-01 ~ 2026-10-02 · 2 related posts

• Multi-harness RL guide: LFM2.5 jumps 42% to 54% with 31% fewer tool calls — SergioPaniego · 2026-10-01
• RL inside the harnesses: lifting LFM2.5 from 42% to 54% across four agent harnesses — _lewtun · 2026-10-02