Robocurve
← Back to trends

All Robocurve runs

LeaderboardAll
Leaderboard: select one or more
BenchmarkAll
Benchmark: select one or more
ModelAll
Model: select one or more
TaskAll
Task: select one or more
SceneAll
Scene: select one or more
EffortAll
Effort: select one or more
Service modeAll
Service mode: select one or more
ControlAll
Control: select one or more
OutcomeAll
Outcome: select one or more

2,109 of 2,109 runs

Robot runs grouped by benchmark, task, and model configuration. Expand groups to view individual trials.
Benchmark / task / model / runScore / outcomeControlEst. costCallsMinutesLogVideoRun data
33.8% mean score
42.9% mean score
23.4% mean score
37.1% mean score
57.8% mean score
5.8% mean score
36.1% mean score
26.4% refused for safety
66.1% refused for safety
0.0% refused for safety
8.3% refused for safety
0.0% refused for safety
25.0% refused for safety
32.8% mean score
62.9% mean score
8.8% mean score
25.8% mean score
RoboDojo-RC Tier 1
  • Zero progress
  • Partial progress
  • Complete
RoboHarm v1
  • Refused (safety)
  • Refused (non-safety)
  • Attempted, failed
  • Harm threshold crossed
  • Provider guardrail block
Simple tasks
  • Zero progress
  • Partial progress
  • Complete