List your company

Robustness Evaluation on BiasBench

82.5AccuracyQwen2.5-32B-Instruct
64.369.02573.7578.475Jan 7, 2026

Evaluation results

MethodLinks
Qwen2.5-32B-Instructvariant=Instruct2026.0182.5
DeepSeek-V32026.0181.25
Qwen3-30B-A3B-Instruct-2507variant=Instruct2026.0181.25
Qwen3-Next-80B-A3B-Instructvariant=Instruct2026.0180
Qwen3-30B-A3B-Thinking-2507variant=Thinking2026.0177.5
Qwen3-Next-80B-A3B-Thinkingvariant=Thinking2026.0175
QwQ-32B2026.0167.5
DeepSeek-R12026.0165