Llama 3.1 Nemotron Nano 4B v1.1

NVIDIA · nvidia/llama-3-1-nemotron-nano-4b-v1-1

ProviderNVIDIA
Familyllama-nemotron
Typellm-chat
Statusactive
Released2025-05-03
Updated2025-10-15
Parameters4.5B
Open weightsyes

Lineage

RelationshipModel
Is a finetune ofnvidia/Llama-3.1-Minitron-4B-Width-Base

What it fits on

Ordered by predicted speed. Bandwidth sets decode rate; memory decides whether it runs at all.

Every figure here is computed, not measured. Fit is weights at each quantisation against device memory, with a 25% allowance for the KV cache, activations and the OS. Decode rate is the memory-bandwidth roofline at 70% efficiency. Nobody has run this model on these devices.
DeviceMemoryBandwidthBest qualityWeightstok/sFastest
Cerebras WSE-344.0 GB21000000.0 GB/sbf169.03 GB~1628719.9~6514879.6 q4
NVIDIA GB200 Grace Blackwell Superchip372.0 GB16000.0 GB/sbf169.03 GB~1240.9~4963.7 q4
AMD Instinct MI355X288.0 GB8000.0 GB/sbf169.03 GB~620.5~2481.9 q4
NVIDIA B200180.0 GB7700.0 GB/sbf169.03 GB~597.2~2388.8 q4
Google TPU7x (Ironwood)192.0 GB7380.0 GB/sbf169.03 GB~572.4~2289.5 q4
AMD Instinct MI325X256.0 GB6000.0 GB/sbf169.03 GB~465.3~1861.4 q4
AMD Instinct MI300X192.0 GB5300.0 GB/sbf169.03 GB~411.1~1644.2 q4
NVIDIA H200 SXM141.0 GB4800.0 GB/sbf169.03 GB~372.3~1489.1 q4
NVIDIA H100 NVL94.0 GB3900.0 GB/sbf169.03 GB~302.5~1209.9 q4
NVIDIA H100 SXM80.0 GB3350.0 GB/sbf169.03 GB~259.8~1039.3 q4
AMD Instinct MI250X128.0 GB3200.0 GB/sbf169.03 GB~248.2~992.7 q4
Google TPU v5p95.0 GB2765.0 GB/sbf169.03 GB~214.4~857.8 q4
NVIDIA A100 80GB SXM80.0 GB2039.0 GB/sbf169.03 GB~158.1~632.6 q4
NVIDIA H100 PCIe80.0 GB2000.0 GB/sbf169.03 GB~155.1~620.5 q4
NVIDIA GeForce RTX 509032.0 GB1792.0 GB/sbf169.03 GB~139.0~555.9 q4
Google TPU v6e (Trillium)32.0 GB1638.0 GB/sbf169.03 GB~127.0~508.2 q4
AMD Instinct MI21064.0 GB1600.0 GB/sbf169.03 GB~124.1~496.4 q4
NVIDIA A100 40GB SXM40.0 GB1555.0 GB/sbf169.03 GB~120.6~482.4 q4
Google TPU v432.0 GB1200.0 GB/sbf169.03 GB~93.1~372.3 q4
NVIDIA GeForce RTX 409024.0 GB1008.0 GB/sbf169.03 GB~78.2~312.7 q4
AMD Radeon RX 7900 XTX24.0 GB960.0 GB/sbf169.03 GB~74.5~297.8 q4
NVIDIA GeForce RTX 508016.0 GB960.0 GB/sbf169.03 GB~74.5~297.8 q4
NVIDIA GeForce RTX 309024.0 GB936.0 GB/sbf169.03 GB~72.6~290.4 q4
NVIDIA L40S48.0 GB864.0 GB/sbf169.03 GB~67.0~268.0 q4
Apple M3 Ultra512.0 GB819.0 GB/sbf169.03 GB~63.5~254.1 q4
AMD Radeon RX 7900 XT20.0 GB800.0 GB/sbf169.03 GB~62.0~248.2 q4
Apple M2 Ultra192.0 GB800.0 GB/sbf169.03 GB~62.0~248.2 q4
Google TPU v5e16.0 GB800.0 GB/sbf169.03 GB~62.0~248.2 q4
NVIDIA GeForce RTX 4080 SUPER16.0 GB736.0 GB/sbf169.03 GB~57.1~228.3 q4
NVIDIA GeForce RTX 4070 Ti SUPER16.0 GB672.0 GB/sbf169.03 GB~52.1~208.5 q4
AMD Radeon RX 9070 XT16.0 GB640.0 GB/sbf169.03 GB~49.6~198.5 q4
Apple M4 Max128.0 GB546.0 GB/sbf169.03 GB~42.3~169.4 q4
Apple M1 Max64.0 GB400.0 GB/sbf169.03 GB~31.0~124.1 q4
Apple M2 Max96.0 GB400.0 GB/sbf169.03 GB~31.0~124.1 q4
Apple M3 Max128.0 GB400.0 GB/sbf169.03 GB~31.0~124.1 q4
NVIDIA GeForce RTX 3060 12GB12.0 GB360.0 GB/sfp84.51 GB~55.8~111.7 q4
NVIDIA L424.0 GB300.0 GB/sbf169.03 GB~23.3~93.1 q4
NVIDIA GeForce RTX 4060 Ti 16GB16.0 GB288.0 GB/sbf169.03 GB~22.3~89.3 q4
Apple M4 Pro64.0 GB273.0 GB/sbf169.03 GB~21.2~84.7 q4
Apple M432.0 GB120.0 GB/sbf169.03 GB~9.3~37.2 q4

Where it runs

Taken from the card's availability section.

PlatformId on that platform
Hugging Facenvidia/Llama-3.1-Nemotron-Nano-4B-v1.1

Verified benchmark evidence

Each row was checked against its source by a reviewer, and carries the model identifier as actually evaluated — which is not always the same as this card's.
BenchmarkEvaluated asScoreEvidence dateSource kind
aa_lcrLlama 3.1 Nemotron Nano 4B v1.10.0%2026-09-10 evaluatedindependent evaluatorsource
gpqa_diamondLlama 3.1 Nemotron Nano 4B v1.140.81%2026-09-10 evaluatedindependent evaluatorsource

Reported benchmark scores

This card reports no benchmark scores yet.

BenchmarkCatalogue standingScore

Data

This card as JSON · See it in the graph · Edit on GitHub