List your company

Reward Modeling on ManiRewardBench Lerobot

0.954Mean VOCTOPReward
-0.03920.218650.47650.73435Feb 22, 2026

Evaluation results

MethodLinks
TOPRewardBackbone=Qwen3-VL-8B2026.020.954
GVLBackbone=Gemini-2.5-Pro, Chat template enforced=true2026.020.62
TOPRewardBackbone=Molmo2-8B2026.020.595
TOPRewardBackbone=Gemini-2.5-Pro, Chat template enforced=true2026.020.578
GVLBackbone=Qwen3-VL-8B2026.020.332
GVLBackbone=Molmo2-8B2026.02-0.001