ByteDance and Tsinghua Open-Source CUDA Agent, Outperforming Human Experts in Code Gen
gekobraa · x · 2026-08-31
ByteDance Seed and Tsinghua University released CUDA Agent, a system using large-scale agentic reinforcement learning. Instead of simple text prompting, the AI iterates code within a sandboxed CUDA environment with automated execution and profiling. Results show it generates higher-performing GPU kernels than human experts and automated compilers like PyTorch, surpassing frontier models like Claude and Gemini.
More from coding & agent
- Vibe-coded dashboard leaked franchise financials before anyone logged in — shauntrennery · 2026-10-16
- Fintech HR built a sourcing agent on 60,000 past applicants — 30 qualified candidates in a month — G0NZ0BOY · 2026-10-16
- Orchestrating a 100-image portrait of humanity for aliens with ChatGPT, Argon and p5.brush — dx8xb · 2026-10-16
- 3x DGX Spark cluster runs GLM-5.3-Flash locally behind a proxy, 1.75M tokens/day at 119W — Teknium · 2026-10-16
- Builder's one rule for trusted desktop agents: read freely, but sending, deleting, buying always ask first — No_Fan5310 · 2026-10-16
- GPT-built unifi-simulator makes MikroTik switches appear inside UniFi Network — Pytlicek · 2026-10-16