Text Retrieval on BEIR-5 (all-MiniLM-L6-v2, test) benchmark leaderboard · SOTA2 Research

Text Retrieval on BEIR-5 (all-MiniLM-L6-v2, test)

52.3Average NDCG@10DoReMi

Evaluation Results

Method	Links
DoReMi2026.01		52.3	32.5	56.3	36.2	48.7	87.7
InfDDS2026.01		52	32.4	58.3	36.1	45.7	87.4
DoReMi2026.01		51.4	31.7	52.3	36.2	49.1	87.6
Static (all-MiniLM-L6-v2)2026.01		51	32.3	51.9	36.9	46.5	87.6
InfDDS2026.01		50.8	32	50.6	34.4	46.6	88
DoGE2026.01		50.5	30	59.6	33.3	43.1	86.4
DoGE2026.01		49.5	29.9	57.4	31.2	42.6	86.3
Static2026.01		49	31.7	49	31.8	44.6	49
CRISP (n=32^3)2026.01		48.5	28.4	61.9	28.5	39.5	84.2
Cooldown2026.01		47.6	28.6	49.9	33.7	38.1	87.6
CRISP (n=32)2026.01		47.4	28.3	58.1	27.9	38	84.9
CRISP (n=32^2)2026.01		45.6	26.8	54.8	26.4	35.5	84.7
MultiDDS2026.01		42.5	22.8	37.2	29.2	28.3	86.3
MultiDDS2026.01		40.2	22.3	43.9	24.8	26.5	83.9