NVIDIA·released 2026-03-071 source
Not yet benchmarked
Claims drawn from cited facts, not live model generation.
Developed by NVIDIA and released in 2026, this model is compact at 3.97 billion parameters. That small scale implies efficiency and lower cost, while offering less headroom than larger, more expensive models.
4 cited facts
This model is not yet benchmarked, with zero tracked benchmarks to its name. It accordingly holds no current top score.
2 cited facts
The average gap to the state of the art is unknown, so position here rests on leaderboard counts rather than a measured margin. It holds the top score on zero benchmarks, which means there is no category leadership yet — not a narrow miss, but no front-of-pack result to point to.
2 cited facts
2 facts cross-checked across data sources: 2 single-source.
Source facts, citations, and refresh stamp for this record.
Sources: huggingface_models
NVIDIA-Nemotron-3-Nano-4B-BF16 is an AI model developed by NVIDIA, released 2026-03-07. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.