Share your thoughts, 1 month free Claude Pro on usSee more

Reward Modeling (Usefulness) on ReasonEdit-Reward-113K 1.0 (test)

0.9428SRCC

RE-Reward

Updated 23d ago

Evaluation Results

Method	Links
RE-Reward 2026.05		0.9428	0.8015	0.9439
Gemini-3-Flash 2026.05		0.8676	0.7364	0.8609
ChatGPT-5.4 2026.05		0.866	0.7467	0.8657
Gemma-4-31B 2026.05		0.8371	0.7078	0.8417
Qwen-3.5-35B-A3B 2026.05		0.7884	0.6729	0.7916
FLEUR 2026.05		0.5393	0.4064	0.5422
Qwen-3.5-9B 2026.05		0.3906	0.3258	0.4037
Qwen3-VL-8B 2026.05		0.3487	0.2907	0.3617
InternVL3-8B 2026.05		0.2841	0.241	0.2346
DeepSeek-VL2-Small 2026.05		0.1848	0.1531	0.1881
Ovis2.5-9B 2026.05		0.1118	0.0747	0.1698
InternVL3.5-8B 2026.05		0.0736	0.0614	0.0291
CLIP-S-ViT-B/32 2026.05		0.0729	0.0501	0.0677
PAC-S++ 2026.05		0.0169	0.0112	0.0181
mPLUG-Owl3-7B 2026.05		0.0023	0.0019	0.003
MiniCPM-V-2.6-8B 2026.05		0	0	0