List your company

Language Modeling on QWEN3-like (val)

2.461LossMACRO-spec
2.438682.589342.742.89066May 6, 2026

Evaluation results

MethodLinks
MACRO-specParameter Scale=1B2026.052.461
MuonH-specParameter Scale=1B2026.052.464
MACRO-froParameter Scale=1B2026.052.467
MuonH-froParameter Scale=1B2026.052.468
MuonParameter Scale=1B2026.052.473
SSOParameter Scale=330M2026.052.712
MACRO-specParameter Scale=330M2026.052.714
MuonH-specParameter Scale=330M2026.052.716
MuonH-froParameter Scale=330M2026.052.717
MACRO-froParameter Scale=330M2026.052.718
FSOParameter Scale=330M2026.052.726
MuonParameter Scale=330M2026.052.736
FSOParameter Scale=120M2026.053.001
MACRO-froParameter Scale=120M2026.053.005
MuonH-froParameter Scale=120M2026.053.007
SSOParameter Scale=120M2026.053.011
MACRO-specParameter Scale=120M2026.053.017
MuonParameter Scale=120M2026.053.019
MuonH-specParameter Scale=120M2026.053.019