List your company

Rational Ordering Evaluation on OR-BIAS-BENCH OOD 200 samples

0.036Biasclaude-haiku-4.5
0.016120.150310.28450.41869Jan 28, 2026

Evaluation results

MethodLinks
claude-haiku-4.52026.010.0360.036—
o4-mini2026.010.0770.01—
Qwen3-8B-Curriculum2026.010.104-0.096—
Qwen3-8B-OM-SFT2026.010.1150.066—
Llama-3.3-70B2026.010.125-0.067—
gpt-4.12026.010.1540.036—
gpt-5-mini2026.010.5330.521—