Reward Modeling on RewardBench Focus 2
90.3AccuracyRubric-ARM-voting@5
Evaluation Results
| Method | Links | |
|---|---|---|
| Rubric-ARM-voting@5voting=52026.02 | 90.3 | |
| Rubric-ARM2026.02 | 89.4 | |
| Claude-3.5-Sonnet2026.02 | 87 | |
| Rubric-RM-voting@5voting=52026.02 | 86.5 | |
| RM-R1-14B (Qwen-2.5-Inst)Backbone=Qwen-2.5-Inst, Size=14B2026.02 | 84.6 | |
| RM-R1-32B (Qwen-2.5-Inst)Backbone=Qwen-2.5-Inst, Size=32B2026.02 | 84.2 | |
| Gemini-2.5-Flash2026.02 | 84.1 | |
| RRM-32BSize=32B2026.02 | 83.6 | |
| Rubric-RM2026.02 | 82.2 | |
| API (Rubric+Judge)Rubric Model=GPT-4.1-Mini, Judge Model=Gemini-2.5-Flash Lite2026.02 | 79.6 | |
| RM-R1-32B (DeepSeek-Dist)Backbone=DeepSeek-Dist, Size=32B2026.02 | 79.2 | |
| RM-R1-14B (DeepSeek-Dist)Backbone=DeepSeek-Dist, Size=14B2026.02 | 79 | |
| RM-R1-7B (Qwen-2.5-Inst)Backbone=Qwen-2.5-Inst, Size=7B2026.02 | 76.2 | |
| Distribution-Calibrated Aggregationn=12, Judge LLM=gemini-2.5-flash2025.12 | 72.5 | |
| FLIPModel Family=QWEN3, Model Size=8B2026.02 | 72.2 | |
| FLIPModel Family=MISTRAL, Model Size=7B2026.02 | 69.3 | |
| Distribution-Calibrated Aggregationn=4, Judge LLM=gemini-2.5-flash2025.12 | 68.5 | |
| FLIPModel Family=QWEN3, Model Size=4B2026.02 | 66.8 | |
| Soft-SCn=12, Judge LLM=gemini-2.5-flash2025.12 | 66.3 | |
| FLIPModel Family=QWEN3, Model Size=0.6B2026.02 | 65.7 | |
| FLIPModel Family=OLMO2, Model Size=7B2026.02 | 65.6 | |
| FLIPModel Family=GEMMA3, Model Size=4B2026.02 | 65.5 | |
| FLIPModel Family=LLAMA3, Model Size=8B2026.02 | 64.1 | |
| Soft-SCn=4, Judge LLM=gemini-2.5-flash2025.12 | 63.6 | |
| API (direct Judge)Judge Model=Gemini-2.5-Flash Lite2026.02 | 63.4 | |
| SCn=4, Judge LLM=gemini-2.5-flash2025.12 | 62.9 | |
| SCn=12, Judge LLM=gemini-2.5-flash2025.12 | 62.6 | |
| CI-SCn=4, Judge LLM=gemini-2.5-flash2025.12 | 61.6 | |
| CI-SCn=12, Judge LLM=gemini-2.5-flash2025.12 | 61.6 | |
| USCn=12, Judge LLM=gemini-2.5-flash2025.12 | 61.2 | |
| RRM-7BSize=7B2026.02 | 60.4 | |
| USCn=4, Judge LLM=gemini-2.5-flash2025.12 | 60.4 | |
| GSCn=12, Judge LLM=gemini-2.5-flash2025.12 | 60.2 | |
| GSCn=4, Judge LLM=gemini-2.5-flash2025.12 | 60.1 | |
| FLIPModel Family=GEMMA3, Model Size=1B2026.02 | 59.1 | |
| FLIPModel Family=QWEN3, Model Size=1.7B2026.02 | 58.6 | |
| Qwen-3-8B (Rubric+Judge)Backbone=Qwen-3-8B2026.02 | 56.6 | |
| RM-R1-7B (DeepSeek-Dist)Backbone=DeepSeek-Dist, Size=7B2026.02 | 55.4 | |
| FLIPModel Family=OLMO2, Model Size=1B2026.02 | 55 | |
| Pairwise Ranking*Model Family=QWEN3, Model Size=4B2026.02 | 51.3 | |
| FLIPModel Family=GEMMA3, Model Size=270M2026.02 | 50.7 | |
| Pairwise Ranking*Model Family=QWEN3, Model Size=8B2026.02 | 48.3 | |
| FLIPModel Family=LLAMA3, Model Size=1B2026.02 | 43.3 | |
| Listwise Ranking*Model Family=QWEN3, Model Size=4B2026.02 | 42.7 | |
| Pointwise RatingModel Family=QWEN3, Model Size=8B2026.02 | 41.3 | |
| Pointwise RatingModel Family=QWEN3, Model Size=4B2026.02 | 40.4 | |
| Pairwise Ranking*Model Family=QWEN3, Model Size=1.7B2026.02 | 39 | |
| Listwise Ranking*Model Family=QWEN3, Model Size=8B2026.02 | 38.5 | |
| FLIPModel Family=LLAMA3, Model Size=3B2026.02 | 38.3 | |
| Listwise Ranking*Model Family=LLAMA3, Model Size=8B2026.02 | 38.1 | |
| Listwise Ranking*Model Family=MISTRAL, Model Size=7B2026.02 | 35.6 | |
| Listwise Ranking*Model Family=QWEN3, Model Size=1.7B2026.02 | 35 | |
| Listwise Ranking*Model Family=GEMMA3, Model Size=4B2026.02 | 34.7 | |
| Pointwise RatingModel Family=QWEN3, Model Size=1.7B2026.02 | 32.4 | |
| JudgeLRM-7BSize=7B2026.02 | 29.1 | |
| Pairwise Ranking*Model Family=GEMMA3, Model Size=4B2026.02 | 28.4 | |
| Pointwise RatingModel Family=GEMMA3, Model Size=4B2026.02 | 26.1 | |
| Listwise Ranking*Model Family=OLMO2, Model Size=7B2026.02 | 26 | |
| Pairwise Ranking*Model Family=MISTRAL, Model Size=7B2026.02 | 25.7 | |
| Pairwise Ranking*Model Family=OLMO2, Model Size=7B2026.02 | 23.2 | |
| Listwise Ranking*Model Family=QWEN3, Model Size=0.6B2026.02 | 22.6 | |
| Listwise Ranking*Model Family=GEMMA3, Model Size=1B2026.02 | 22 | |
| Pairwise Ranking*Model Family=GEMMA3, Model Size=1B2026.02 | 21.8 | |
| Listwise Ranking*Model Family=LLAMA3, Model Size=3B2026.02 | 21 | |
| Pairwise Ranking*Model Family=LLAMA3, Model Size=8B2026.02 | 20.5 | |
| Pairwise Ranking*Model Family=QWEN3, Model Size=0.6B2026.02 | 20.1 | |
| Pointwise RatingModel Family=QWEN3, Model Size=0.6B2026.02 | 18.8 | |
| Pointwise RatingModel Family=GEMMA3, Model Size=270M2026.02 | 15.3 | |
| Pointwise RatingModel Family=LLAMA3, Model Size=8B2026.02 | 14.9 | |
| Listwise Ranking*Model Family=OLMO2, Model Size=1B2026.02 | 13.7 | |
| Pointwise RatingModel Family=OLMO2, Model Size=1B2026.02 | 13.6 | |
| Pairwise Ranking*Model Family=GEMMA3, Model Size=270M2026.02 | 13.4 | |
| Listwise Ranking*Model Family=LLAMA3, Model Size=1B2026.02 | 13.2 | |
| Pairwise Ranking*Model Family=LLAMA3, Model Size=3B2026.02 | 12.8 | |
| Listwise Ranking*Model Family=GEMMA3, Model Size=270M2026.02 | 12.4 | |
| Pointwise RatingModel Family=GEMMA3, Model Size=1B2026.02 | 12 | |
| Pointwise RatingModel Family=LLAMA3, Model Size=1B2026.02 | 11.5 | |
| Pointwise RatingModel Family=LLAMA3, Model Size=3B2026.02 | 9.9 | |
| Pointwise RatingModel Family=MISTRAL, Model Size=7B2026.02 | 9.7 | |
| Pairwise Ranking*Model Family=LLAMA3, Model Size=1B2026.02 | 7.8 | |
| Pointwise RatingModel Family=OLMO2, Model Size=7B2026.02 | 7.4 | |
| Pairwise Ranking*Model Family=OLMO2, Model Size=1B2026.02 | 2.1 |