List your company

Language Understanding on MMLU (Accuracy and Inference Time)

73.1AccuracyUnwatermarked
55.31659.93364.5569.167May 28, 2026

Evaluation results

MethodLinks
UnwatermarkedBase Model=Ministral3-8B2026.0573.11
UnwatermarkedBase Model=Qwen3-8B2026.0572.61
De-markBase Model=Qwen3-8B2026.0572.49.47
Watermarked BaselineBase Model=Qwen3-8B2026.0572.31
WASHBase Model=Qwen3-8B, n=42026.0572.31.01
WASHBase Model=Qwen3-8B, n=52026.0572.31.01
ToBlendBase Model=Qwen3-8B2026.0572.21.01
WASHBase Model=Qwen3-8B, n=32026.0572.21.01
WASHBase Model=Ministral3-8B, n=42026.0571.81.03
WASHBase Model=Ministral3-8B, n=52026.0571.81.03
De-markBase Model=Ministral3-8B2026.0571.710.02
Watermarked BaselineBase Model=Ministral3-8B2026.0571.61
ToBlendBase Model=Ministral3-8B2026.0571.61.04
WASHBase Model=Ministral3-8B, n=32026.0571.61.04
WASHBase Model=Llama3.1-8B, n=52026.0564.61.06
WASHBase Model=Llama3.1-8B, n=42026.0564.41.06
ToBlendBase Model=Llama3.1-8B2026.0563.11.06
WASHBase Model=Llama3.1-8B, n=32026.0563.11.06
UnwatermarkedBase Model=Llama3.1-8B2026.0558.31
De-markBase Model=Llama3.1-8B2026.0556.5—
Watermarked BaselineBase Model=Llama3.1-8B2026.05561