List your company

Instruction Following on Inverse IFEval, MARS-Bench, MultiChallenge, Collie-Hard, EIFBench

99Pass@1GPT-5 High
42.52857.18971.8586.511Mar 21, 2026

Evaluation results

MethodLinks
GPT-5 HighDataset=Collie-Hard2026.0399
Gemini 3-proDataset=Collie-Hard2026.0395.5
Gemini 3-proDataset=MARS-Bench2026.0380.8
Gemini 3-proDataset=Inverse IFEval2026.0380.6
Seed1.8Dataset=Inverse IFEval2026.0380.3
GPT-5 HighDataset=Inverse IFEval2026.0378.9
Claude Sonnet-4.5Dataset=Collie-Hard2026.0377.6
GPT-5 HighDataset=MARS-Bench2026.0377.2
Gemini 2.5-proDataset=Inverse IFEval2026.0375.3
Gemini 2.5-proDataset=MARS-Bench2026.0373.6
Seed1.8Dataset=Collie-Hard2026.0372.6
Claude Sonnet-4.5Dataset=MARS-Bench2026.0372.5
Claude Sonnet-4.5Dataset=Inverse IFEval2026.0370.2
Seed1.8Dataset=MARS-Bench2026.0370.1
GPT-5 HighDataset=MultiChallenge2026.0369.6
Gemini 2.5-proDataset=Collie-Hard2026.0369.5
Gemini 3-proDataset=MultiChallenge2026.0367.4
Seed1.8Dataset=MultiChallenge2026.0366.7
GPT-5 HighDataset=EIFBench2026.0366.7
Claude Sonnet-4.5Dataset=MultiChallenge2026.0357.2
Gemini 2.5-proDataset=MultiChallenge2026.0355.4
Gemini 3-proDataset=EIFBench2026.0350.1
Seed1.8Dataset=EIFBench2026.0348.6
Claude Sonnet-4.5Dataset=EIFBench2026.0347
Gemini 2.5-proDataset=EIFBench2026.0344.7