{
  "benchmark": "RealBench",
  "suite": "System-hard",
  "model": "GPT6.1-sol",
  "source": "user_reported_results",
  "source_description": "用户提供的 RealBench System-hard 结果表",
  "runs_per_task": null,
  "time_budget": null,
  "token_budget": null,
  "tool_versions": null,
  "configurations": [
    {"id": "ours", "name": "我们的 Agent", "label": "Deep Valley Chip", "reported_label": "GPT6.1-sol+ours"},
    {"id": "codex", "name": "Codex", "label": "Codex", "reported_label": "GPT6.1-sol+codex"},
    {"id": "claude", "name": "Claude Code", "label": "Claude Code", "reported_label": "GPT6.1-sol+Claude code"}
  ],
  "tasks": [
    {"id": "aes_cipher_top", "label": "AES 加密", "results": {"ours": "pass", "codex": "pass", "claude": "pass"}},
    {"id": "aes_inv_cipher_top", "label": "AES 解密", "results": {"ours": "pass", "codex": "pass", "claude": "pass"}},
    {"id": "sdc_controller", "label": "SDC 控制器", "results": {"ours": "pass", "codex": "fail", "claude": "fail"}},
    {"id": "e203_exu_alu", "label": "E203 执行单元", "results": {"ours": "pass", "codex": "pass", "claude": "pass"}},
    {"id": "e203_ifu", "label": "E203 取指单元", "results": {"ours": "pass", "codex": "pass", "claude": "fail"}}
  ]
}
