NGen-4 Model Cards
TNSA
company
Verified
AI & ML interests
Building India's first Foundational Model and AGI
Recent Activity
TNSA's OpenWeight Model distilled from NGen4 Foundational Models "Note : We recommend you to use the new NGen-4-OW Family instead of this"
TNSA's NGen4 Family this Collection Includes Experimental Model Builds, these Models are the experimental Architectures and Experiments done.
-
TNSA/NGen4-Base-Exp-a-05-08-2025-pre
Text Generation • 51M • Updated • 2 • 1 -
TNSA/NGen4-Base-Exp-b-05-08-2025-pre
Text Generation • 51M • Updated • 2 • 1 -
TNSA/NGen4-Base-Exp-c-05-08-2025-pre
Text Generation • 51M • Updated • 1 • 1 -
TNSA/NGen4-Nano-Exp-a-05-09-2025-pre
Text Generation • Updated • 2 • 1
Datasets for Training of High-Level AI Systems, and few Extractions from them.
Datasets for Training Models
The NGen-4-OW models are built on top of distillation data of NGen-4 Pro Model by TNSA running on NGen-4-OW Architecture.
TNSA AI's NGen 3 Model Squad the Architecture is not Open Source, only the weights are Open please visit, https://tnsaai.com to read System Card
Benchmark dataset for Benching
Best Reasoning Dataset in One Place
All Models from the NGen2 Model Family, Pre-Trained for Token-Prediction, are experimental and not made for Production use-case. Check out NGen-4-OW.
NGen-4 Model Cards
The NGen-4-OW models are built on top of distillation data of NGen-4 Pro Model by TNSA running on NGen-4-OW Architecture.
TNSA's OpenWeight Model distilled from NGen4 Foundational Models "Note : We recommend you to use the new NGen-4-OW Family instead of this"
TNSA AI's NGen 3 Model Squad the Architecture is not Open Source, only the weights are Open please visit, https://tnsaai.com to read System Card
TNSA's NGen4 Family this Collection Includes Experimental Model Builds, these Models are the experimental Architectures and Experiments done.
-
TNSA/NGen4-Base-Exp-a-05-08-2025-pre
Text Generation • 51M • Updated • 2 • 1 -
TNSA/NGen4-Base-Exp-b-05-08-2025-pre
Text Generation • 51M • Updated • 2 • 1 -
TNSA/NGen4-Base-Exp-c-05-08-2025-pre
Text Generation • 51M • Updated • 1 • 1 -
TNSA/NGen4-Nano-Exp-a-05-09-2025-pre
Text Generation • Updated • 2 • 1
Benchmark dataset for Benching
Datasets for Training of High-Level AI Systems, and few Extractions from them.
Best Reasoning Dataset in One Place
Datasets for Training Models
All Models from the NGen2 Model Family, Pre-Trained for Token-Prediction, are experimental and not made for Production use-case. Check out NGen-4-OW.