ByteDance and Tsinghua train an RL agent to write faster GPU kernels

ByteDance Seed and Tsinghua AIR released CUDA Agent, a large-scale agentic reinforcement learning (RL) system that trains a language model to write GPU kernels that outperform standard compiler output. The target gap is narrow: frontier models already write correct CUDA code, they just write slow CUDA, and the base model (Seed1.6) already passes 74.0% of tasks correctly on KernelBench before RL training.

Why it matters: GPU kernel optimization is a narrow but high-leverage bottleneck — small per-kernel speedups compound across massive training and inference runs, directly cutting compute costs at scale. Using RL specifically to close the 'correct but slow' gap points toward AI increasingly optimizing the low-level infrastructure that trains and serves other AI models, not just user-facing application code.