Nvidia is not just one of the few companies that still use Llama as a base model, it is also one of the few American model providers with a cadence and quality that matches Chinese companies, as this release proves yet again. Interestingly, it is way less censored than its predecessor as evaluated by SpeechMap:
Specs
Params49B
LicenseNVIDIA Open Model License
Tags
modelsmodels/llmsmodels/llms/reasoningmultilingualartifacts/13
RAM score
—
Hugging Face Downloads
86.2K
last 30d
1.9M
all time
HF Likes
234
Relative Adoption Metric contextualizes downloads against the model's size bucket.
VAIL Fingerprint
00b5:0158:0291:03a2:079d:1144:2663:624b
Explore other models with behavioral similarity to Llama-3.3-Nemotron-Super-49B-v1.5.
Related Models
