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

Logical Reasoning on SynLogic

8.7pass@1

F-GRPO

4.1245.3126.57.688Feb 6, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.02
8.757
2026.02
8.260
2026.02
853.6
2026.02
7.951.3
2026.02
7.962.4
2026.02
7.553.3
2026.02
5.433.9
2026.02
5.226.3
2026.02
526.9
2026.02
4.927.4
2026.02
4.926.4
2026.02
4.926.1
2026.02
4.827.5
2026.02
4.828.9
2026.02
4.736.4
2026.02
4.635.5
2026.02
4.333
2026.02
4.329.4