[
 {
  "group": "Ranked, simulation and real",
  "name": "RoboDojo",
  "sim": "ARX X5",
  "real": "ARX X5, Piper, Piper X",
  "verdict": "same-name",
  "detail": "Same team and policies, trained separately on sim and real demos. 7 task names in both tracks, not aligned by design; 2 on the same robot.",
  "used": "70 cells",
  "source": "arXiv 2607.04434 Table 2; robodojo-benchmark.com",
  "finetune": "Separate weights: sim demos for sim, real demos for real"
 },
 {
  "group": "Ranked, real robots",
  "name": "RoboDojo-RC",
  "sim": "–",
  "real": "I2RT YAM",
  "verdict": "replica",
  "detail": "Rebuilds six of RoboDojo's real tasks on YAM arms with adapted rubrics. Only GR00T N1.7 is tested on both, too few to compare.",
  "used": "–",
  "source": "github.com/robocurve/robodojo-rc",
  "finetune": "–"
 },
 {
  "group": "Ranked, real robots",
  "name": "DataCenterBench",
  "sim": "–",
  "real": "I2RT YAM",
  "verdict": "no",
  "detail": "Real only.",
  "used": "–",
  "source": "robocurve.org/datacenterbench",
  "finetune": "–"
 },
 {
  "group": "Ranked, real robots",
  "name": "AA-Robotics",
  "sim": "–",
  "real": "Hiwonder NexArm",
  "verdict": "no",
  "detail": "Real only.",
  "used": "–",
  "source": "artificialanalysis.ai",
  "finetune": "–"
 },
 {
  "group": "Ranked, real robots",
  "name": "ArmnetBench",
  "sim": "–",
  "real": "SO-101",
  "verdict": "no",
  "detail": "Real only.",
  "used": "–",
  "source": "arXiv 2607.24481",
  "finetune": "–"
 },
 {
  "group": "Ranked, real robots",
  "name": "VLA-Replica",
  "sim": "–",
  "real": "SO-101, two labs",
  "verdict": "replica",
  "detail": "No simulator. The real setup was rebuilt in a second lab and four policies rerun.",
  "used": "20 cells, real-to-real reference",
  "source": "arXiv 2605.20774 Table 4, App. G",
  "finetune": "–"
 },
 {
  "group": "Ranked, real robots",
  "name": "RoboChallenge",
  "sim": "–",
  "real": "UR5, Franka, ARX5, ALOHA",
  "verdict": "no",
  "detail": "Real only.",
  "used": "–",
  "source": "arXiv 2510.17950",
  "finetune": "–"
 },
 {
  "group": "Ranked, real robots",
  "name": "RoboArena",
  "sim": "–",
  "real": "Franka (DROID)",
  "verdict": "no",
  "detail": "Real only.",
  "used": "–",
  "source": "arXiv 2506.18123",
  "finetune": "–"
 },
 {
  "group": "Ranked, real robots",
  "name": "PhAIL",
  "sim": "–",
  "real": "Franka FR3 (DROID)",
  "verdict": "no",
  "detail": "Real only.",
  "used": "–",
  "source": "arXiv 2605.29710",
  "finetune": "–"
 },
 {
  "group": "Ranked, real robots",
  "name": "RoboColosseum (open and fine-tune tracks)",
  "sim": "–",
  "real": "Franka (DROID), two-arm YAM, SO-101, Unitree G1",
  "verdict": "no",
  "detail": "Real only: its leaderboard has no simulation track. Not AgiBot’s RoboColiseum.",
  "used": "–",
  "source": "robocolosseum.ai",
  "finetune": "–"
 },
 {
  "group": "Ranked, simulation",
  "name": "RoboTwin 2.0",
  "sim": "ALOHA-AgileX",
  "real": "sim-trained policies deployed",
  "verdict": "no",
  "detail": "Real experiments test policy transfer; the real-tested checkpoints have no sim scores.",
  "used": "–",
  "source": "arXiv 2506.18088 Table 4",
  "finetune": "–"
 },
 {
  "group": "Ranked, simulation",
  "name": "RoboLab-120",
  "sim": "Franka (DROID)",
  "real": "4 policies, v1 and v2 only",
  "verdict": "same-name",
  "detail": "Matched sim and real success on six unnamed simple tasks for four policies, as one aggregate. In v1 and v2; removed from v3 without comment and replaced by a comparison with RoboArena. Cited as v2, which never changes.",
  "used": "4 results",
  "source": "arXiv 2604.09860v2 Table V",
  "finetune": "None (DROID-trained policies)"
 },
 {
  "group": "Ranked, simulation",
  "name": "MolmoSpaces",
  "sim": "Franka FR3",
  "real": "Franka (DROID), MolmoBot paper",
  "verdict": "same-name",
  "detail": "Same checkpoints in sim and real on related pick-and-place tasks; scenes not mirrored.",
  "used": "12 cells",
  "source": "arXiv 2603.16861v2 Tables 6, 7",
  "finetune": "None: sim-trained MolmoBot runs zero-shot on the real robot"
 },
 {
  "group": "Ranked, simulation",
  "name": "RoboCasa365",
  "sim": "mobile Franka",
  "real": "sim-trained policies deployed",
  "verdict": "no",
  "detail": "Policy transfer only.",
  "used": "–",
  "source": "arXiv 2603.04356 Table 5",
  "finetune": "–"
 },
 {
  "group": "Ranked, simulation",
  "name": "RoboColiseum (Genie Sim 3.0)",
  "sim": "Agibot G1 and G2",
  "real": "Agibot G1 (platform paper only)",
  "verdict": "twin",
  "detail": "Its platform paper scores four pi0.5 checkpoints per task in mirrored sim and real scenes on an Agibot G1. RoboColiseum's own leaderboard runs on the G2 and has no real counterpart.",
  "used": "16 cells",
  "source": "arXiv 2601.02078v4 Table I",
  "finetune": "Some checkpoints trained on sim data, some on real"
 },
 {
  "group": "Ranked, simulation",
  "name": "EBench",
  "sim": "Isaac Sim arms",
  "real": "–",
  "verdict": "no",
  "detail": "Sim only; sim-real correlation left to future work.",
  "used": "–",
  "source": "arXiv 2606.18239",
  "finetune": "–"
 },
 {
  "group": "Ranked, simulation",
  "name": "LIBERO-Pro2.0",
  "sim": "Franka",
  "real": "–",
  "verdict": "no",
  "detail": "Sim only.",
  "used": "–",
  "source": "arXiv 2610.24587",
  "finetune": "–"
 },
 {
  "group": "Ranked, simulation",
  "name": "LIBERO-Plus",
  "sim": "Franka",
  "real": "–",
  "verdict": "no",
  "detail": "Sim only.",
  "used": "–",
  "source": "arXiv 2510.13626",
  "finetune": "–"
 },
 {
  "group": "Ranked, simulation",
  "name": "Embodied First Exam",
  "sim": "44 robots",
  "real": "–",
  "verdict": "no",
  "detail": "Sim only; its sim-to-real suite deploys to a second simulator.",
  "used": "–",
  "source": "embodied-first-exam.ai",
  "finetune": "–"
 },
 {
  "group": "Ranked, paper-reported",
  "name": "SimplerEnv",
  "sim": "Google Robot, WidowX",
  "real": "Google Robot, WidowX",
  "verdict": "twin",
  "detail": "Sim scenes built to copy the real setups; same checkpoints in both. Visual matching is the twin; variant aggregation is used as an unmatched control.",
  "used": "92 cells",
  "source": "arXiv 2405.05941 Tables IV, V",
  "finetune": "None (real-trained policies)"
 },
 {
  "group": "Ranked, paper-reported",
  "name": "LIBERO",
  "sim": "Franka",
  "real": "–",
  "verdict": "no",
  "detail": "Sim only.",
  "used": "–",
  "source": "arXiv 2306.03310",
  "finetune": "–"
 },
 {
  "group": "Ranked, paper-reported",
  "name": "CALVIN",
  "sim": "Franka",
  "real": "–",
  "verdict": "no",
  "detail": "Sim only.",
  "used": "–",
  "source": "arXiv 2112.03227",
  "finetune": "–"
 },
 {
  "group": "Collected, not ranked",
  "name": "PolaRiS",
  "sim": "splat replicas",
  "real": "Franka (DROID)",
  "verdict": "twin",
  "detail": "Six real scenes reconstructed as Gaussian splats. Per-cell sim scores are after sim co-fine-tuning; the paper reports r = 0.30 without it. Its LIBERO scores serve as an unmatched control.",
  "used": "60 cells",
  "source": "arXiv 2512.16881; polaris-evals.github.io data",
  "finetune": "Yes: about 1k steps on sim demos from other scenes"
 },
 {
  "group": "Collected, not ranked",
  "name": "BEHAVIOR",
  "sim": "Galaxea R1 Pro",
  "real": "–",
  "verdict": "no",
  "detail": "The challenge is sim only. A BEHAVIOR-1K pilot has one approximate paired result, too small to use.",
  "used": "–",
  "source": "arXiv 2403.09227 App. G",
  "finetune": "–"
 },
 {
  "group": "Collected, not ranked",
  "name": "RoboChallenge v2",
  "sim": "–",
  "real": "five robots",
  "verdict": "no",
  "detail": "Real only.",
  "used": "–",
  "source": "RoboChallenge dataset card",
  "finetune": "–"
 },
 {
  "group": "Collected, not ranked",
  "name": "LIBERO-Pro",
  "sim": "Franka",
  "real": "–",
  "verdict": "no",
  "detail": "Sim only.",
  "used": "–",
  "source": "arXiv 2510.03827",
  "finetune": "–"
 },
 {
  "group": "Collected, not ranked",
  "name": "Embodied Analysis",
  "sim": "unnamed two-arm suite",
  "real": "–",
  "verdict": "no",
  "detail": "Sim only, and it does not name its tasks.",
  "used": "–",
  "source": "embodiedanalysis.com",
  "finetune": "–"
 }
]