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

Reasoning and Knowledge Suite (MMLU, ARC, BoolQ, CSQA, HSwag, PIQA, SocIQ, Wino)

75.78MMLU

Qwen3-4B

36.218446.489256.7667.0308Dec 8, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.12
75.7889.8397.5386.0981.979.4684.9875.5965.4381.84
2025.12
66.8686.4492.5983.8876.0973.8581.4569.463.6977.14
2025.12
65.3580.3491.8979.8274.6160.7677.268.5859.2773.09
2025.12
63.0477.2988.5476.1270.5269.279.0565.2564.472.6
2025.12
61.5679.3290.4876.3975.164.1876.1764.9459.6771.98
2025.12
57.8772.283.9576.7370.3571.0679.0564.3364.0971.07
2025.12
56.3670.1783.671.970.5260.7775.7363.4659.8368.04
2025.12
55.266.4482.5472.4269.4566.278.8965.9265.3569.16
2025.12
55.0968.1484.4869.0561.1848.5169.9761.5155.6463.73
2025.12
53.966.182.8978.5367.458.1374.3762.5965.7567.74
2025.12
51.9959.6682.7269.8567.1665.378.5160.1859.1266.05
2025.12
51.7664.7582.5478.5966.1861.277.3163.2561.8867.5
2025.12
44.2547.4676.7270.5565.661.6176.4455.5360.3862.06
2025.12
37.7436.9570.5567.4362.8260.274.9250.6158.1757.71