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

Reward Modeling on LitBench

80.7Accuracy

AC-GenRM

Updated 10d ago

Evaluation Results

Method	Links
AC-GenRM 2026.04		80.7
AC-GenRM 2026.04		79.6
AC-GenRM 2026.04		77.6
claude-3-7-sonnet-20250219 2026.01		76.31
gpt-5-2025-08-07 2026.01		76.17
RM-NLHF-Qwen-32B 2026.01		74.92
Claude-Sonnet-3.7 2026.04		73.1
AC-GenRM 2026.04		72.8
o3-2025-04-16 2026.01		72.02
gemini-2.5-pro 2026.01		71.81
ArmoRM-Llama3-8B-v0.1 2026.01		71.57
INF-ORM-Llama3.1-70B 2026.01		71.41
DeepSeek-R1 2026.04		71
GPT-4.1 2026.04		70.2
deepseek-r1-0528 2026.01		70
DeepSeek-V3 2026.04		70
RM-R1-Qwen-32B 2026.01		69.34
Base 2026.04		68.8
qwen3-max 2026.01		68.59
qwen-plus-latest 2026.01		68.35
Claude-3.5-Haiku 2026.04		67.5
RRM-Qwen-32B 2026.01		67.46
Base 2026.04		66.6
Skywork-Reward-Llama-3.1-8B-v0.2 2026.01		65.93
RM-NLHF-Qwen-7B 2026.01		65.83
deepseek-v3 2026.01		65.36
gpt-4o-latest 2026.01		63.19
URM-LLaMa-3.1-8B 2026.01		63.02
GPT-4.1-Mini 2026.04		63
R1-Distill-Qwen-32B 2026.01		62.06
R1-Distill-Llama-8B 2026.01		58.22
RM-R1-Qwen-7B 2026.01		57.5
Base 2026.04		54.3
DeepSeek-R1-Distill-Qwen-7B 2026.01		54.18
RRM-Qwen-7B 2026.01		53.83
Base 2026.04		47.7