Our new X account is live! Follow @wizwand_team for updates

Reward Modeling on RM-Bench (test)

87.1Overall Score

Qwen3-30B-A3B

Updated 4d ago

Evaluation Results

Method	Links
Qwen3-30B-A3B 2026.02		87.1	74.9	95.5	84.4	93.6	84.6
Qwen3-14B 2026.02		86.7	75.7	95.7	82.2	93.2	82.9
Principles-Qwen32B 2026.02		86.2	80.4	92	77	95.5	83.8
GRAM-R² 2026.02		85.7	76	89.8	80.6	96.2	83.4
Qwen3-30B-A3B (Outcome-Only) 2026.02		84.9	68.7	95.9	81.3	93.6	80.3
RM-R1-Distilled-Qwen-32B 2026.02		83.9	74.2	91.8	74.1	95.4	81.4
Qwen3-14B (Outcome-Only) 2026.02		83.6	72.3	92.6	77.8	91.7	76.8
CE-RM-4B 2026.01		83.2	-	-	-	-	-
RewardAnything-8B-v1 2026.02		83.1	76.7	90.3	75.2	90.2	72.9
RRM-32B 2026.01		82.8	-	-	-	-	-
Nemotron-Super 2026.02		82.7	73.7	91.4	75	90.6	80
J1-Llama-70B 2026.01		82.7	-	-	-	-	-
CE-RM-4B 2026.01		81.9	-	-	-	-	-
RM-R1-Distilled-Qwen-14B 2026.02		81.5	71.8	90.5	69.5	94.1	79.8
Think-J-32B 2026.01		79.8	-	-	-	-	-
CE-RM-4B 2026.01		79.8	-	-	-	-	-
RM-R1-Qwen-32B 2026.01		79.1	-	-	-	-	-
TIR-Judge-Distill-8B 2026.01		76.7	-	-	-	-	-
TIR-Judge-Zero-8B 2026.01		76.3	-	-	-	-	-
J1-Llama-8B 2026.01		73.4	-	-	-	-	-
CompassJudger2-32B 2026.01		73.2	-	-	-	-	-
RRM-32B 2026.02		73.1	66.6	81.4	65.2	79.4	74.4
DeepSeek-R1-0528 2026.02		72.8	76.7	74.3	51	89.2	75.8
TIR-Judge-Zero-4B 2026.01		72.8	-	-	-	-	-
GPT-4o 2026.02		72.5	67.2	67.5	63.6	91.7	66.2
GPT-4o 2026.01		72.5	-	-	-	-	-
RISE-Judge-32B 2026.01		72.3	-	-	-	-	-
TIR-Judge-Distill-4B 2026.01		71.9	-	-	-	-	-
CompassJudger2-7B 2026.01		71.5	-	-	-	-	-
RRM-7B 2026.01		70.8	-	-	-	-	-
Gemini-2.5-Flash 2026.01		70.8	-	-	-	-	-
Skywork-Reward-Gemma-2-27B 2026.02		70.5	71.8	59.2	56.6	94.3	67.8
RM-R1-Qwen-7B 2026.01		70.2	-	-	-	-	-
Skywork-Reward-Llama-3.1-8B 2026.02		70.1	69.5	60.6	54.5	95.7	66.3
Llama-3.3-70B-Instruct 2026.01		69.5	-	-	-	-	-
RISE-Judge-7B 2026.01		68	-	-	-	-	-
Claude-3.5-Sonnet 2026.02		61	62.5	62.6	54.4	64.4	62.9
JudgeLRM-7B 2026.01		58.7	-	-	-	-	-
CompassJudger1-32B 2026.01		54.4	-	-	-	-	-