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

Reward Modeling on RM-Bench

96Accuracy

Skywork-Reward-V2-Llama-3.1-8B-40M

Updated 1d ago

Evaluation Results

Method
Skywork-Reward-V2-Llama-3.1-8B-40M 2025.07	96	-	-	-	-
Gemini-3-Flash 2026.03	93.8	-	-	-	-
OpenRS 2026.02	93	-	-	-	89.4
Skywork-Reward-V2-Llama-3.1-8B 2026.02	92.8	-	-	-	84.3
Skywork-Reward-V2-Llama-3.1-8B 2025.07	92.8	-	-	-	-
OpenRS 2026.02	92.7	-	-	-	86.7
OpenRS 2026.02	91.8	-	-	-	86.2
OpenRS 2026.02	91.7	-	-	-	86.8
DeepSeek-V3.2 2026.03	91.4	-	-	-	-
OpenRS 2026.02	91.1	-	-	-	84.9
RationaleRM 2026.02	87.1	-	-	-	-
CAMEL-Reflection 2026.02	84.2	-	-	-	-
RM-R1-DeepSeek-Distill-Qwen-32B 2026.02	83.9	-	-	-	-
RM-R1-DeepSeek-Distill-Qwen-32B 2025.07	83.9	-	-	-	-
RM-R1-DEEPSEEK-DISTILLED-QWEN-32B 2025.05	83.9	-	-	-	-
J1-Llama-70B 2026.02	82.7	-	-	-	-
J1-Llama-70B 2026.02	82.7	-	-	-	-
Mix-GRM 2026.03	82.7	-	-	-	-
J1-Llama-70B 2025.07	82.7	-	-	-	-
Skywork-Reward-V2-Qwen3-8B 2025.07	82.6	-	-	-	-
SAVE 2026.05	82.3	-	-	-	-
EvalPlanner 2026.02	82.1	-	-	-	-
EvalPlanner 2025.07	82.1	-	-	-	-
SAVE 2026.05	82.1	-	-	-	-
SAVE (w/o Curriculum Mechanism) 2026.05	82	-	-	-	-
SAVE (w/o Policy Model Optimization) 2026.05	82	-	-	-	-
CAMEL 2026.02	81.9	-	-	-	-
SAVE (w/o Curriculum Mechanism) 2026.05	81.9	-	-	-	-
SAVE (w/o Policy Model Optimization) 2026.05	81.7	-	-	-	-
Skywork-Reward-V2-Qwen3-4B 2025.07	81.6	-	-	-	-
RM-R1-DEEPSEEK-DISTILLED-QWEN-14B 2025.05	81.5	-	-	-	-
HL-BT 2026.05	81.4	-	-	-	-
Skywork-Reward-V2-Llama-3.2-3B 2025.07	81.1	-	-	-	-
HL-BT 2026.05	81	-	-	-	-
Continual Offline Training RM 2026.05	80.6	-	-	-	-
EvalPlanner 2025.07	80	-	-	-	-
Skywork-Reward-V2-Llama-3.2-3B 2026.05	80	-	-	-	-
Mix-GRM 2026.03	79.2	-	-	-	-
RM-R1-Qwen-Instruct-32B 2026.02	79.1	-	-	-	-
RM-R1-Qwen-Instruct-32B 2025.07	79.1	-	-	-	-
RM-R1-QWEN-INSTRUCT-32B 2025.05	79.1	-	-	-	-
Base-GRM 2026.03	78.8	-	-	-	-
Skywork-Reward-V2-Qwen3-1.7B 2025.07	78.7	-	-	-	-
JudgeLRM-7B 2026.03	78.5	-	-	-	-
ConsistRM (Qwen3-8B) 2026.04	78.3	-	-	-	-
Qwen3-8B + RFT 2026.04	78.2	-	-	-	-
Qwen3-8B + TTRL 2026.04	77.4	-	-	-	-
Qwen3-8B + CAAR 2026.04	77.3	-	-	-	-
Qwen3-8B + SFT 2026.04	77.1	-	-	-	-
Base-GRM 2026.03	77	-	-	-	-
ConsistRM (Qwen3-4B) 2026.04	76.8	-	-	-	-
RM-R1-7B (Distill) 2026.03	76.6	-	-	-	-
Qwen3-4B + RFT 2026.04	76.6	-	-	-	-
Qwen3-4B + CAAR 2026.04	76.6	-	-	-	-
Mean Reward 2026.05	76.6	-	-	-	-
Skywork-Reward-V2-Llama-3.2-1B 2025.07	76.3	-	-	-	-
Qwen3-4B + TTRL 2026.04	76.3	-	-	-	-
RM-R1-QWEN-INSTRUCT-14B 2025.05	76.1	-	-	-	-
Qwen3-4B + SFT 2026.04	76.1	-	-	-	-
Qwen3-8B 2026.04	75.8	-	-	-	-
Mean Reward 2026.05	75.8	-	-	-	-
INF-ORM-Llama3.1-70B 2025.07	75.4	-	-	-	-
Gemini-1.5-pro 2025.05	75.2	-	-	-	-
RM-R1-7B (Instruct) 2026.03	75.1	-	-	-	-
CAMEL-Fast 2026.02	74.8	-	-	-	-
Claude-3.5-Sonnet 2025.07	74.5	-	-	-	-
Qwen3-4B 2026.04	74.5	-	-	-	-
Skywork-Reward-V2-Qwen3-0.6B 2025.07	74.4	-	-	-	-
FARE-8B 2026.03	74.1	-	-	-	-
RubricRM-8B 2026.03	74	-	-	-	-
INF-ORM-Llama3.1-70B 2026.02	73.8	-	-	-	71.7
INF-ORM-Llama3.1-70B 2026.02	73.8	-	-	-	-
J1-Llama-8B 2026.02	73.4	-	-	-	-
J1-Llama-8B 2025.07	73.4	-	-	-	-
J1-Llama-8B 2026.04	73.4	-	-	-	-
RM-R1-Qwen-Instruct-32B 2026.02	73.1	-	-	-	-
GPT-4o 2025.07	73.1	-	-	-	-
QRM-Llama3.1-8B 2025.07	72.8	-	-	-	-
GPT-4o 2026.02	72.5	-	-	-	-
RM-R1-Qwen-Instruct-14B 2026.02	72.5	-	-	-	-
QRM-Llama3.1-8B-v2 2025.07	72.5	-	-	-	-
GPT-4o-0806 2025.05	72.5	-	-	-	-
Skywork-Reward-8B 2026.03	72.4	-	-	-	-
RM-R1-DEEPSEEK-DISTILLED-QWEN-7B 2025.05	72.4	-	-	-	-
Llama-3.1-Nemotron-70B 2026.02	72.2	-	-	-	70.5
Llama-3.1-Nemotron-70B 2026.02	72.2	-	-	-	-
Llama-3.1-Nemotron-70B 2025.07	72.2	-	-	-	-
Skywork-Reward-Llama-3.1-8B-v0.2 2026.02	72.1	-	-	-	67.1
Skywork-Reward-Llama-3.1-8B-v0.2 2026.02	72.1	-	-	-	-
Internlm2-20b-reward 2025.07	72.1	-	-	-	-
URM-LLaMa-3.1-8B 2025.07	72	-	-	-	-
Skywork-Critic-Llama-3.1-70B 2025.05	71.9	-	-	-	-
Skywork-Reward-Llama-3.1-8B 2025.07	71.8	-	-	-	-
Self-taught-evaluator-llama3.1-70B 2025.05	71.4	-	-	-	-
Llama-3-OffsetBias-RM-8B 2026.02	71.3	-	-	-	65
Llama-3-OffsetBias-RM-8B 2026.02	71.3	-	-	-	-
Llama-3-OffsetBias-RM-8B 2025.07	71.3	-	-	-	-
LDL-Reward-Gemma-2-27B-v0.1 2026.02	71.1	-	-	-	67.4
LDL-Reward-Gemma-2-27B-v0.1 2026.02	71.1	-	-	-	-
LDL-Reward-Gemma-2-27B-v0.1 2025.07	71	-	-	-	-

Showing 100 of 226 rows