NVIDIA fine-tuned Llama models and turned them into the Llama Nemotron suite of open models.
Mostly true True in substance, wrong in a detail.
Factual: Statements about the present or the past, checked against the record: did they hold up?
The evidence is the person's or their organisation's own records.
Said for an organisation they lead or work for; it counts on both records.
What would count as true or false
Judged as of 2025-01-07. True: NVIDIA announced (on or before that date) a Llama Nemotron family of open models built by fine-tuning or otherwise post-training Meta's Llama models. Mostly true: the family was announced as described but differs in a minor detail (e.g. availability came later than the announcement, or extra techniques such as pruning or distillation were used besides fine-tuning). Mixed: only some of the models are Llama-derived or open. Mostly false: the models are not openly available, or not mainly derived from Llama. False: no such NVIDIA Llama Nemotron suite existed or was announced.
Set before the research began.
Evidence (4)
-
E1 NVIDIA Announces Nemotron Model Families to Advance Agentic AI NVIDIA (NVIDIA blog) · 7 January 2025 Own record
“NVIDIA today announced the Llama Nemotron family of open large language models (LLMs). Built with Llama, the models can help developers create and deploy AI agents”
-
E2 NVIDIA Announces Nemotron Model Families to Advance Agentic AI NVIDIA (NVIDIA blog) · 7 January 2025 Own record
“The NVIDIA Llama Nemotron models use NVIDIA NeMo for distilling, pruning and alignment. Using these techniques, the models are small enough to run on a variety of computing platforms”
-
E3 NVIDIA Announces Nemotron Model Families to Advance Agentic AI NVIDIA (NVIDIA blog) · 7 January 2025 Own record
“Llama Nemotron and Cosmos Nemotron models will be available soon as hosted application programming interfaces and for download on build.nvidia.com and Hugging Face.”
-
E4 NVIDIA Launches AI Foundation Models for RTX AI PCs NVIDIA Corporation (GLOBE NEWSWIRE press release) · 6 January 2025 Own record
“NVIDIA today also announced the Llama Nemotron family of open models that provide high accuracy on a wide range of agentic tasks.”
Reasoning
NVIDIA's official blog and its press release announced the Llama Nemotron family of open models, "built with Llama", on 6–7 January 2025 (E1, E4). The same announcement says NeMo was used for distilling, pruning and alignment, not only fine-tuning (E2). It also says the models would be available "soon", so they could not yet be downloaded on the announcement date (E3). Under the criteria, these minor differences make the claim mostly true. The verdict rests only on NVIDIA's own official announcements.
Cite this claim
Jensen Huang, “NVIDIA fine-tuned Llama models and turned them into the Llama Nemotron suite of open models.” (said 7 January 2025). Verdict: mostly true. Reliability API, claim c_318462749705, method 0.1.0+03d23c0e, dataset release v1.1.2. https://epistacklabs.com/claims/c_318462749705
This claim is in dataset release v1.1.2, with the same verdict.
Reply to this claim
More factual claims by Jensen Huang
-
Mixed
“For the first time, anybody in the world can program a computer.”
For the first time, anybody in the world can program a computer, using human language.
-
Mostly true
“The NVIDIA Isaac GR00T is our humanoid robotic stack model, data generation, simulation, the runtime, including the operating system.”
-
Mixed
“Frontier smart, five times faster, 30% cheaper, completely open.”
-
Resolved true
“Today we’re announcing the NVIDIA Isaac GR00T, a reference humanoid robot, all fully integrated.”
-
Mostly true
“The longer you think, the better the quality answer you get.”
AI models that think longer (inference-time reasoning) produce better quality answers.
Verdicts are made automatically; how verdicts are made.