NVIDIA · nvidia/nvidia-nemotron-3-nano-4b-bf16
| Provider | NVIDIA |
|---|---|
| Family | nemotron |
| Type | llm-chat |
| Status | active |
| Released | 2026-03-07 |
| Updated | 2026-03-20 |
| Parameters | 4B |
| Open weights | yes |
| Relationship | Model |
|---|---|
| Is a finetune of | nvidia/NVIDIA-Nemotron-Nano-9B-v2 |
Ordered by predicted speed. Bandwidth sets decode rate; memory decides whether it runs at all.
| Device | Memory | Bandwidth | Best quality | Weights | tok/s | Fastest |
|---|---|---|---|---|---|---|
| Cerebras WSE-3 | 44.0 GB | 21000000.0 GB/s | bf16 | 7.95 GB | ~1849728.2 | ~7398912.7 q4 |
| NVIDIA GB200 Grace Blackwell Superchip | 372.0 GB | 16000.0 GB/s | bf16 | 7.95 GB | ~1409.3 | ~5637.3 q4 |
| AMD Instinct MI355X | 288.0 GB | 8000.0 GB/s | bf16 | 7.95 GB | ~704.7 | ~2818.6 q4 |
| NVIDIA B200 | 180.0 GB | 7700.0 GB/s | bf16 | 7.95 GB | ~678.2 | ~2712.9 q4 |
| Google TPU7x (Ironwood) | 192.0 GB | 7380.0 GB/s | bf16 | 7.95 GB | ~650.0 | ~2600.2 q4 |
| AMD Instinct MI325X | 256.0 GB | 6000.0 GB/s | bf16 | 7.95 GB | ~528.5 | ~2114.0 q4 |
| AMD Instinct MI300X | 192.0 GB | 5300.0 GB/s | bf16 | 7.95 GB | ~466.8 | ~1867.3 q4 |
| NVIDIA H200 SXM | 141.0 GB | 4800.0 GB/s | bf16 | 7.95 GB | ~422.8 | ~1691.2 q4 |
| NVIDIA H100 NVL | 94.0 GB | 3900.0 GB/s | bf16 | 7.95 GB | ~343.5 | ~1374.1 q4 |
| NVIDIA H100 SXM | 80.0 GB | 3350.0 GB/s | bf16 | 7.95 GB | ~295.1 | ~1180.3 q4 |
| AMD Instinct MI250X | 128.0 GB | 3200.0 GB/s | bf16 | 7.95 GB | ~281.9 | ~1127.5 q4 |
| Google TPU v5p | 95.0 GB | 2765.0 GB/s | bf16 | 7.95 GB | ~243.5 | ~974.2 q4 |
| NVIDIA A100 80GB SXM | 80.0 GB | 2039.0 GB/s | bf16 | 7.95 GB | ~179.6 | ~718.4 q4 |
| NVIDIA H100 PCIe | 80.0 GB | 2000.0 GB/s | bf16 | 7.95 GB | ~176.2 | ~704.7 q4 |
| NVIDIA GeForce RTX 5090 | 32.0 GB | 1792.0 GB/s | bf16 | 7.95 GB | ~157.8 | ~631.4 q4 |
| Google TPU v6e (Trillium) | 32.0 GB | 1638.0 GB/s | bf16 | 7.95 GB | ~144.3 | ~577.1 q4 |
| AMD Instinct MI210 | 64.0 GB | 1600.0 GB/s | bf16 | 7.95 GB | ~140.9 | ~563.7 q4 |
| NVIDIA A100 40GB SXM | 40.0 GB | 1555.0 GB/s | bf16 | 7.95 GB | ~137.0 | ~547.9 q4 |
| Google TPU v4 | 32.0 GB | 1200.0 GB/s | bf16 | 7.95 GB | ~105.7 | ~422.8 q4 |
| NVIDIA GeForce RTX 4090 | 24.0 GB | 1008.0 GB/s | bf16 | 7.95 GB | ~88.8 | ~355.1 q4 |
| AMD Radeon RX 7900 XTX | 24.0 GB | 960.0 GB/s | bf16 | 7.95 GB | ~84.6 | ~338.2 q4 |
| NVIDIA GeForce RTX 5080 | 16.0 GB | 960.0 GB/s | bf16 | 7.95 GB | ~84.6 | ~338.2 q4 |
| NVIDIA GeForce RTX 3090 | 24.0 GB | 936.0 GB/s | bf16 | 7.95 GB | ~82.4 | ~329.8 q4 |
| NVIDIA L40S | 48.0 GB | 864.0 GB/s | bf16 | 7.95 GB | ~76.1 | ~304.4 q4 |
| Apple M3 Ultra | 512.0 GB | 819.0 GB/s | bf16 | 7.95 GB | ~72.1 | ~288.6 q4 |
| AMD Radeon RX 7900 XT | 20.0 GB | 800.0 GB/s | bf16 | 7.95 GB | ~70.5 | ~281.9 q4 |
| Apple M2 Ultra | 192.0 GB | 800.0 GB/s | bf16 | 7.95 GB | ~70.5 | ~281.9 q4 |
| Google TPU v5e | 16.0 GB | 800.0 GB/s | bf16 | 7.95 GB | ~70.5 | ~281.9 q4 |
| NVIDIA GeForce RTX 4080 SUPER | 16.0 GB | 736.0 GB/s | bf16 | 7.95 GB | ~64.8 | ~259.3 q4 |
| NVIDIA GeForce RTX 4070 Ti SUPER | 16.0 GB | 672.0 GB/s | bf16 | 7.95 GB | ~59.2 | ~236.8 q4 |
| AMD Radeon RX 9070 XT | 16.0 GB | 640.0 GB/s | bf16 | 7.95 GB | ~56.4 | ~225.5 q4 |
| Apple M4 Max | 128.0 GB | 546.0 GB/s | bf16 | 7.95 GB | ~48.1 | ~192.4 q4 |
| Apple M1 Max | 64.0 GB | 400.0 GB/s | bf16 | 7.95 GB | ~35.2 | ~140.9 q4 |
| Apple M2 Max | 96.0 GB | 400.0 GB/s | bf16 | 7.95 GB | ~35.2 | ~140.9 q4 |
| Apple M3 Max | 128.0 GB | 400.0 GB/s | bf16 | 7.95 GB | ~35.2 | ~140.9 q4 |
| NVIDIA GeForce RTX 3060 12GB | 12.0 GB | 360.0 GB/s | bf16 | 7.95 GB | ~31.7 | ~126.8 q4 |
| NVIDIA L4 | 24.0 GB | 300.0 GB/s | bf16 | 7.95 GB | ~26.4 | ~105.7 q4 |
| NVIDIA GeForce RTX 4060 Ti 16GB | 16.0 GB | 288.0 GB/s | bf16 | 7.95 GB | ~25.4 | ~101.5 q4 |
| Apple M4 Pro | 64.0 GB | 273.0 GB/s | bf16 | 7.95 GB | ~24.0 | ~96.2 q4 |
| Apple M4 | 32.0 GB | 120.0 GB/s | bf16 | 7.95 GB | ~10.6 | ~42.3 q4 |
Taken from the card's availability section.
| Platform | Id on that platform |
|---|---|
| Hugging Face | nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 |
| Benchmark | Evaluated as | Score | Evidence date | Source kind | |
|---|---|---|---|---|---|
| aa_lcr | Nemotron 3 Nano 4B | 21.67% | 2026-09-10 evaluated | independent evaluator | source |
| critpt | Nemotron 3 Nano 4B | 0.0% | 2026-09-10 evaluated | independent evaluator | source |
| gdpval_aa | Nemotron 3 Nano 4B | 0.0% | 2026-09-10 evaluated | independent evaluator | source |
| gpqa_diamond | Nemotron 3 Nano 4B | 51.31% | 2026-09-10 evaluated | independent evaluator | source |
This card reports no benchmark scores yet.
| Benchmark | Catalogue standing | Score |
|---|