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

Search on MM Search (Pass@4)

61Pass@4

AXPO

Updated 6d ago

Evaluation Results

Method	Links
AXPO 2026.05		61
GRPO 2026.05		60
SFT 2026.05		59.7
SFT + GRPO 2026.05		59.3
Base 2026.05		59.3
Base 2026.05		59
AXPO 2026.05		57.3
SFT + GRPO 2026.05		57
AXPO 2026.05		57
SFT + GRPO 2026.05		55.7
Base 2026.05		55.7
SFT 2026.05		55.3
GRPO 2026.05		54.7
SFT 2026.05		53.7
Base 2026.05		46.3
GRPO 2026.05		42