Multi-objective preference selection on OpenAI models Setup A4: K=5, M*=5 benchmark leaderboard · SOTA2 Research

Multi-objective preference selection on OpenAI models Setup A4: K=5, M*=5

0.052Cost ($)Cost-constrained LMA

Evaluation Results

Method	Links
Cost-constrained LMA2026.05		0.052	3	80
CUPID without language2026.05		0.058	3.4	100
CUPID2026.05		0.063	2.4	100
LMA2026.05		0.073	3.5	80
ϵ-greedy2026.05		0.075	2	80
CUPID without cost penalty2026.05		0.091	3	100
RUCB2026.05		0.118	1	33.3
Latent Dueling Bandits2026.05		0.122	6.75	80