Skip to content
Updated minutes ago
NVIDIA

Nemotron 15B

NVIDIA · dense · 15B parameters · 4,096 context

Quality
72.0

Parameters

15B

Context Window

4K tokens

Architecture

Dense

Best GPU

H100 SXM

Cheapest API

$0.30/M

Quality Score

72/100

Intelligence Brief

Nemotron 15B is a 15B parameter DENSE model from NVIDIA, featuring Grouped Query Attention (GQA) with 32 layers and 6,144 hidden dimensions. With a 4,096 token context window, it supports tools, structured output, code, math, multilingual. On standardized benchmarks, it achieves MMLU 71, HumanEval 40, GSM8K 75. The most cost-effective API deployment is via nvidia at $0.30/M output tokens. For self-hosted inference, H100 SXM delivers optimal throughput at $1794/month.

Architecture Details

TypeDENSE
Total Parameters15B
Active Parameters15B
Layers32
Hidden Dimension6,144
Attention Heads48
KV Heads8
Head Dimension128
Vocab Size256,000

Memory Requirements

BF16 Weights

30.0 GB

FP8 Weights

15.0 GB

INT4 Weights

7.5 GB

KV-Cache per Token131072 bytes
Activation Estimate1.50 GB

GPU Compatibility Matrix

Nemotron 15B is compatible with 82% of GPU configurations across 41 GPUs at 3 precision levels.

BF16 (Full)
FP8 (Half)
INT4 (Quarter)
Blackwell(7 GPUs)
B200 NVL (pair)360GB
B300288GB
B100 SXM192GB
GB200 NVL72 (per GPU)192GB
Hopper(7 GPUs)
H100 NVL 94GB (per GPU pair)188GB
H200 SXM141GB
H2096GB
GH20096GB
Ada Lovelace(11 GPUs)
L40S48GB
L4048GB
RTX 6000 Ada48GB
L2048GB
Ampere(16 GPUs)
A100 80GB SXM80GB
A100 80GB PCIe80GB
A1664GB
RTX A600048GB
Legend:No fitVery tightTightModerateGoodExcellent

GPU Recommendations

H100 SXMoptimal

FP8 · 1 GPU · tensorrt-llm

100/100

score

Throughput

1.1K tok/s

Latency (ITL)

1.0ms

Est. TTFT

0ms

Cost/Month

$1794

Cost/M Tokens

$0.65

Use this config →
H100 PCIeoptimal

FP8 · 1 GPU · tensorrt-llm

95/100

score

Throughput

1.1K tok/s

Latency (ITL)

1.0ms

Est. TTFT

0ms

Cost/Month

$1794

Cost/M Tokens

$0.65

Use this config →
RTX A6000optimal

BF16 · 1 GPU · vllm

95/100

score

Throughput

138.2 tok/s

Latency (ITL)

7.2ms

Est. TTFT

1ms

Cost/Month

$465

Cost/M Tokens

$1.28

Use this config →

Deployment Options

API

API Deployment

nvidia

$0.30/M

output tokens

Self-Hosted

Single GPU

H100 SXM

$1794/mo

Min VRAM: 15 GB

Scale

Multi-GPU

RTX 3090 x2

280.4 tok/s

TP· $361/mo

API Pricing Comparison

ProviderInput $/MOutput $/MBadges
nvidia$0.30$0.30
Cheapest

Cost Analysis

ProviderInput $/MOutput $/M~Monthly Cost
nvidiaBest Value$0.30$0.30$3

Cost per 1,000 Requests

Short (500 tok)

$0.21

via nvidia

Medium (2K tok)

$0.84

via nvidia

Long (8K tok)

$3.00

via nvidia

Performance Estimates

Throughput by GPU

H100 SXM
1.1K tok/s
H100 PCIe
1.1K tok/s
RTX A6000
138.2 tok/s

VRAM Breakdown (H100 SXM, FP8)

Weights
Act
Weights 15.0 GBKV-Cache 1.1 GBActivations 12.0 GBOverhead 0.8 GB

Precision Impact

bf16

30.0 GB

weights/GPU

fp8

15.0 GB

weights/GPU

~1.1K tok/s

int4

7.5 GB

weights/GPU

Quality Benchmarks

Above Average
76th percentile across all models
MMLU
71.0
Below Average (34th pctile)
HumanEval
40.0
Below Average (25th pctile)
GSM8K
75.0
Below Average (36th pctile)
MT-Bench
76.0
Bottom 25% (0th pctile)

Capabilities

Features

Tool Use Vision Code Math Reasoning Multilingual Structured Output

Supported Frameworks

vllmsglangtensorrt-llm

Supported Precisions

BF16 (default)FP8INT4

Where to Deploy Nemotron 15B

Similar Models

Minitron 8B

8B params · dense

Quality: 62

from $0.10/M

Larger context, Lower quality, CheaperCompare →

Qwen 2.5 14B

14.8B params · dense

Quality: 76

from $0.30/M

Larger contextCompare →
Larger context, Higher qualityCompare →

Phi-4

14.7B params · dense

Quality: 73

from $0.14/M

Larger context, CheaperCompare →
Larger context, Lower qualityCompare →

Frequently Asked Questions

How much VRAM does Nemotron 15B need for inference?

Nemotron 15B requires approximately 30.0 GB of VRAM at BF16 precision, 15.0 GB at FP8, or 7.5 GB at INT4 quantization. Additional VRAM is needed for KV-cache (131072 bytes per token) and activations (~1.50 GB).

What is the best GPU for Nemotron 15B?

The top recommended GPU for Nemotron 15B is the H100 SXM using FP8 precision. It achieves approximately 1.1K tokens/sec at an estimated cost of $1794/month ($0.65/M tokens). Score: 100/100.

How much does Nemotron 15B inference cost?

Nemotron 15B API inference starts from $0.30/M input tokens and $0.30/M output tokens. Self-hosted inference costs depend on your GPU configuration — use our ROI calculator for a detailed breakdown.