List your company

General Capability on All Benchmarks Overall

52.04Overall Average ScoreUltraMix
30.220835.885441.5547.2146Jul 21, 2025Aug 9, 2025Aug 28, 2025Sep 17, 2025Oct 6, 2025Oct 25, 2025Nov 14, 2025

Evaluation results

MethodLinks
UltraMixTraining Method=DPO, Dataset=UM-190k2025.1152.04
UltraMixTraining Method=DPO, Dataset=UM-187k2025.1151.74
TuluDPOTraining Method=DPO, Dataset=TuluDPO2025.1150.87
UltraMixTraining Method=DPO, Dataset=UM-170k2025.1150.55
ORPOTraining Method=ORPO, Dataset=ORPO2025.1149.55
DataAgentRLModel Backbone=LLaMA-DCLM, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0746.84
SmolLM-3-3B-SFTTraining Method=SFT2025.1146.61
HelpSteerTraining Method=DPO, Dataset=HelpSteer2025.1146.51
CodePrefTraining Method=DPO, Dataset=CodePref2025.1145.92
DataAgentRLModel Backbone=LLaMA-Nemo, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0745.9
DataAgentSFTModel Backbone=LLaMA-DCLM, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0745.75
UltraFBTraining Method=DPO, Dataset=UltraFB2025.1145.69
DataAgentRLModel Backbone=LLaMA-FWE, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0745.55
DataAgentSFTModel Backbone=LLaMA-FWE, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0745.21
DataAgentSFTModel Backbone=LLaMA-Nemo, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0745.16
RegMixModel Backbone=LLaMA-Nemo, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0744.86
RegMixModel Backbone=LLaMA-DCLM, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0744.67
RegMixModel Backbone=LLaMA-FWE, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0743.78
Naive TrainingModel Backbone=LLaMA-DCLM, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0738.51
Naive TrainingModel Backbone=LLaMA-FWE, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0738.51
Base ModelModel Backbone=LLaMA-Nemo, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0738.22
Base ModelModel Backbone=LLaMA-DCLM, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0738.15
Naive TrainingModel Backbone=LLaMA-Nemo, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0737.86
Base ModelModel Backbone=LLaMA-FWE, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0737.65
DataAgentRLModel Backbone=Pythia-1.4B, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0735.25
RegMixModel Backbone=Pythia-1.4B, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0734.64
DataAgentSFTModel Backbone=Pythia-1.4B, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0734.01
Base ModelModel Backbone=Pythia-1.4B, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0733.47
Naive TrainingModel Backbone=Pythia-1.4B, Data Reweighting Space=52-dimensional (Nvidia domain classifier)2025.0731.06