General Instruction Following on Arena-Hard v2

85.9Score

Updated 1mo ago

Evaluation Results

Method	Links
o3 2026.02		85.9
Gemini-2.5-Pro 2026.02		79
Qwen3-30B-A3B-Instruct + More Query Rubrics 2026.02		74.6
GenRM-R-Align-14B 2026.02		60.2
Claude-3.7-Sonnet 2026.02		59.8
GenRM-R-Align-8B 2026.02		59.5
Deepseek-R1 2026.02		56.8
GenRM-RLVR-14B 2026.02		55.9
GenRM-RLVR-8B 2026.02		53.1
Qwen3-14B-as-GenRM 2026.02		51
GPT-4.1 2026.02		50
Qwen3-235B-Instruct 2026.02		46.7
Qwen3-8B-as-GenRM 2026.02		46.6
Baichuan-M2-32B 2026.02		45.8
Qwen3-32B 2026.02		44.5
HuatuoGPT-o1-72B 2026.02		43.2
Qwen3-4B-Instruct + More Query Rubrics 2026.02		41.2
Qwen3-4B-Instruct + Doctor Rubrics 2026.02		39.7
Qwen3-4B-Instruct + Principle Rubrics 2026.02		37
Qwen3-4B-Instruct + Draft Rubrics 2026.02		34.9
Qwen3-30B-A3B-Instruct 2026.02		33.9
Qwen3-8B 2026.02		26.5
+RL (Skywork-Reward-V2-Llama-3.1-8B) 2025.07		18.5
+RL (Skywork-Reward-V2-Qwen3-4B) 2025.07		17.9
Instruct (official) 2025.07		17.1
+RL (Skywork-Reward-Gemma-2-27B-v0.2) 2025.07		15.5
Qwen3-4B-Instruct 2026.02		15
+RL (Skywork-Reward-Llama-3-8B-v0.2) 2025.07		12.2
+SFT 2025.07		9.9
+RL (Skywork-Reward-V2-Llama-3.1-8B) 2025.07		6.3
+RL (Skywork-Reward-V2-Qwen3-4B) 2025.07		6
Instruct (official) 2025.07		5.8
Base 2025.07		5.6
+RL (Skywork-Reward-Gemma-2-27B-v0.2) 2025.07		3.8
+SFT 2025.07		3.1
Base 2025.07		2
+RL (Skywork-Reward-Llama-3-8B-v0.2) 2025.07		1.6