List your company

Instruction Following Evaluation on FollowBench OOD

78.06HSRPrompt
21.48436.17250.8665.548Jun 22, 2026

Evaluation results

MethodLinks
PromptModel=Deepseek-V4-Flash, Category=Vanilla2026.0678.0684.62
PromptModel=Qwen3-32B, Category=Vanilla2026.0664.6677.12
EvoRubricsModel=Qwen3-8B, Category=Ours2026.0663.0174.7
OnlineRubricsModel=Qwen3-8B, Category=RL w/ Rubric2026.0660.7374.9
SFTModel=Qwen3-8B, Category=Supervised2026.0659.3372.4
GoldenRubricsModel=Qwen3-8B, Category=RL w/ Rubric2026.0658.3170.61
EvoRubricsModel=Qwen3-4B, Category=Ours2026.0653.9266.2
PromptModel=Qwen3-4B, Category=Vanilla2026.0647.7664.09
RuscaRLModel=Qwen3-8B, Category=RL w/ Rubric2026.0639.3461.59
SFTModel=Qwen3-4B, Category=Supervised2026.0638.7656.31
GoldenRubricsModel=Qwen3-4B, Category=RL w/ Rubric2026.0632.550.4
RuscaRLModel=Qwen3-4B, Category=RL w/ Rubric2026.0631.8952.1
PromptModel=Qwen3-8B, Category=Vanilla2026.0623.8936.49
OnlineRubricsModel=Qwen3-4B, Category=RL w/ Rubric2026.0623.6641.47