NVIDIA-Nemotron-3-Nano-4B

Nemotron family · 4.0B params

Evidence: 1 sourced run · 2 awaiting review · 1 independent source · first recorded Jul 20, 2026 · last recorded Aug 25, 2026. What these labels mean.

What the NVIDIA-Nemotron-3-Nano-4B numbers mean

The fastest verified result on our board is 185.8 output tok/s on an RTX 5070 at Q4_K_M with a 32768-token context. That is about 5.4 ms per generated token — 500 tokens takes roughly 2.7 seconds. Prompt processing on the same record runs at 5377.5 tok/s, so a 2,000-token prompt is read in about 0.4 seconds. Peak VRAM was measured at 4.2 GB. Across the 3 records we hold for this model the range is 171.3–237.8 tok/s on 3 GPUs, because quantisation, context length and the engine build all move the number.

How much VRAM does NVIDIA-Nemotron-3-Nano-4B need?

NVIDIA-Nemotron-3-Nano-4B has 4.0B parameters. These estimates use the same formula as the VRAM calculator: weights, plus about 15% for a moderate-context KV cache and runtime overhead, plus a small engine floor.

Long contexts and large batches raise the real number — read the smallest card as "possible with a reduced context". Measured throughput is on the benchmark board.

Models with a similar footprint

Ranked by distance from NVIDIA-Nemotron-3-Nano-4B's parameter count. Full benchmark board.

All benchmarks Methodology