Agents, and RL environment, for optimizing GPU kernels on AMD ROCm using LLM agents. Benchmarks LLM serving workloads end-to-end, profiles bottleneck kernels, optimizes them via Claude Code or Codex, and scores on compilation, correctness, and speedup.
Latest release v0.1.0 · 7 Aug 2026
2 files for Codex, OpenCode and Claude Code: Apex AGENTS.md, Apex CLAUDE.md — 19,966 tokens loaded in every session.
AGENTS.md A 9,941 tok CLAUDE.md A 10,025 tok These files are AMD-AGI/Apex's own configuration — they tell Codex, OpenCode and Claude Code how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.