Language Modeling on Open LLM Leaderboard Unified General Ability Suite (test) benchmark leaderboard · SOTA2 Research

Language Modeling on Open LLM Leaderboard Unified General Ability Suite (test)

12MMLU-P AccuracyHyperGLU

Evaluation Results

Method	Links
HyperGLU2026.02		12	26.8	32.1	1.2	46.8	28.6	18.9	36.6	61.1	51.6	70.8	61.3	53.1	76	35.3	78.8	1.88	9
HyperMLP2026.02		11.9	26.2	30	1	46.6	27.3	15.5	28.8	54.1	40.2	66.3	58.5	51.3	71	32	75.5	2.97	4
HyperGLU2026.02		11.7	25.9	31.3	0.9	45.1	27.3	15.7	30.4	53.8	42.7	67.4	58.7	51.8	72	33.6	77.1	2.31	6
HGRN22026.02		11.5	25.4	29.5	0.2	35	22.3	12.9	28.2	50.4	31.7	67.1	41.6	52.2	77	32.8	37.8	5.84	1
HGRN2026.02		11.4	26.9	29.7	0.8	40.9	25.3	12.2	27.1	51.8	48.1	70.7	58.4	51.5	70	32.6	69.5	5.12	0
SoftmaxAttn2026.02		11.4	25.9	29.6	1.1	36.5	27	14.1	28	49.2	49.2	70.5	62.1	55.2	76	31.8	76.9	4.22	2
GLA2026.02		11.4	25.9	29.5	0.6	42.7	27.2	15.7	27.7	48.2	48.8	70.2	57.4	54.1	69	32.6	72.1	5.19	0
GatedDeltaNet2026.02		11.3	26	29.6	1	42.1	25.8	13.3	27.6	52.7	39.6	66.2	58.8	52.7	71	33.8	73.5	3.81	2
DeltaNet2026.02		11.2	26	30	0.9	45.2	27.7	15	26.9	50.2	40.5	65.3	51.9	50.4	69	31.6	71.7	4.81	1
ReLUAttn+KVConv2026.02		11.1	25.3	29	0.8	42.3	22.9	13	26.8	52.5	41.2	66.9	52.4	52	66	32.4	77.4	5.13	1
GSA2026.02		11	27	29.4	1.3	43.8	30	17.9	28.7	52.9	51	71.2	54.1	53.6	76	33	77.3	2.84	4
RetNet2026.02		11	25.2	29.3	0.1	38.4	5.6	2.4	27.1	48.9	48	70.1	58.3	53.3	71	32.4	73.6	6.69	0
GLA2026.02		11	25.8	28.9	0.7	41.5	22.8	10.9	24.7	47.8	36.6	63.7	54.7	48.9	64	29.4	64.9	7.56	0
DeltaNet2026.02		10.9	26.3	30.8	1.1	41.7	28.8	16.5	26.6	52.2	50.2	70.4	61.1	54.1	74	31.8	76.1	4.22	0
DiffTrans2026.02		10.9	25.9	29.9	0.8	39	26.6	13.3	28.9	53.1	40.8	66.8	60.3	53.4	69	33	73.4	4.13	2
SoftmaxAttn2026.02		10.6	26.7	29.2	1	38.6	26.9	12.6	27.3	50.6	39.6	65	56.9	49.9	72	32.4	72.7	5.28	3