DaisyRose-27B boosts Qwen3.8 performance on Terminal-Bench 4.1
The DaisyRose-27B model, optimized through reinforcement learning, increases Qwen3.8 performance from 5.6% to 12% on Terminal-Bench 4.1.
Published 28m1 source
Lire en français
≈ 10s
12 %
6% to 12% on Terminal-Bench 4
Click the link to read an article on the topic: