Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Reranking on Hard-case evaluation set Opus-4.6

91.2MAP

Gemini-3-Flash

78.92882.11485.388.486May 7, 2026
Updated 26d ago

Evaluation Results

MethodLinks
2026.05
91.297.895.293.897.454.882.268.582.2
2026.05
9197.895.193.397.354.282.867.782.8
2026.05
89.398.595.592.196.85478.867.578.8
2026.05
86.29793.689.69650.476.66376.6
2026.05
8495.592.787.295.248.474.660.574.6
2026.05
79.791.58582.492.945.6705770
2026.05
79.492.686.98393.445.26956.569