Skip to content
Updated minutes ago
Stability

StableLM 2 12B

Stability AI · dense · 12.1B parameters · 4,096 context

Quality
50.0

Parameters

12.1B

Context Window

4K tokens

Architecture

Dense

Best GPU

A100 40GB SXM

Cheapest API

$0.25/M

Intelligence Brief

StableLM 2 12B is a 12.1B parameter DENSE model from Stability AI, featuring Grouped Query Attention (GQA) with 40 layers and 5,120 hidden dimensions. With a 4,096 token context window, it supports code, multilingual. The most cost-effective API deployment is via stabilityai at $0.25/M output tokens. For self-hosted inference, A100 40GB SXM delivers optimal throughput at $807/month.

Architecture Details

TypeDENSE
Total Parameters12.1B
Active Parameters12.1B
Layers40
Hidden Dimension5,120
Attention Heads32
KV Heads8
Head Dimension160
Vocab Size100,352

Memory Requirements

BF16 Weights

24.2 GB

FP8 Weights

12.1 GB

INT4 Weights

6.0 GB

KV-Cache per Token204800 bytes
Activation Estimate1.00 GB

GPU Compatibility Matrix

StableLM 2 12B 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

A100 40GB SXMoptimal

BF16 · 1 GPU · vllm

95/100

score

Throughput

312.3 tok/s

Latency (ITL)

3.2ms

Est. TTFT

1ms

Cost/Month

$807

Cost/M Tokens

$0.98

Use this config →
RTX A6000optimal

BF16 · 1 GPU · vllm

95/100

score

Throughput

154.2 tok/s

Latency (ITL)

6.5ms

Est. TTFT

1ms

Cost/Month

$465

Cost/M Tokens

$1.15

Use this config →
A40optimal

BF16 · 1 GPU · vllm

95/100

score

Throughput

139.8 tok/s

Latency (ITL)

7.2ms

Est. TTFT

1ms

Cost/Month

$399

Cost/M Tokens

$1.09

Use this config →

Deployment Options

API

API Deployment

stabilityai

$0.25/M

output tokens

Self-Hosted

Single GPU

A100 40GB SXM

$807/mo

Min VRAM: 12 GB

Scale

Multi-GPU

RTX 3090 x2

307.0 tok/s

TP· $361/mo

API Pricing Comparison

ProviderInput $/MOutput $/MBadges
stabilityai$0.25$0.25
Cheapest

Cost Analysis

ProviderInput $/MOutput $/M~Monthly Cost
stabilityaiBest Value$0.25$0.25$3

Cost per 1,000 Requests

Short (500 tok)

$0.17

via stabilityai

Medium (2K tok)

$0.70

via stabilityai

Long (8K tok)

$2.50

via stabilityai

Performance Estimates

Throughput by GPU

A100 40GB SXM
312.3 tok/s
RTX A6000
154.2 tok/s
A40
139.8 tok/s

VRAM Breakdown (A100 40GB SXM, BF16)

Weights
Act
Weights 24.2 GBKV-Cache 3.4 GBActivations 8.0 GBOverhead 1.9 GB

Precision Impact

bf16

24.2 GB

weights/GPU

~312.3 tok/s

fp8

12.1 GB

weights/GPU

int4

6.0 GB

weights/GPU

Capabilities

Features

Tool Use Vision Code Math Reasoning Multilingual Structured Output

Supported Frameworks

vllmsglangtgiollama

Supported Precisions

BF16 (default)FP8INT4

Where to Deploy StableLM 2 12B

Similar Models

Larger contextCompare →

Gemma 3 12B

12B params · dense

Quality: 71

from $0.10/M

Larger context, Higher quality, CheaperCompare →
Larger context, Higher quality, CheaperCompare →

Pixtral 12B

12B params · dense

Quality: 50

from $0.15/M

Larger context, CheaperCompare →

FLUX.1 Dev

12B params · dense

Quality: 50

from $25.00/M

Smaller context, More expensiveCompare →

Frequently Asked Questions

How much VRAM does StableLM 2 12B need for inference?

StableLM 2 12B requires approximately 24.2 GB of VRAM at BF16 precision, 12.1 GB at FP8, or 6.0 GB at INT4 quantization. Additional VRAM is needed for KV-cache (204800 bytes per token) and activations (~1.00 GB).

What is the best GPU for StableLM 2 12B?

The top recommended GPU for StableLM 2 12B is the A100 40GB SXM using BF16 precision. It achieves approximately 312.3 tokens/sec at an estimated cost of $807/month ($0.98/M tokens). Score: 95/100.

How much does StableLM 2 12B inference cost?

StableLM 2 12B API inference starts from $0.25/M input tokens and $0.25/M output tokens. Self-hosted inference costs depend on your GPU configuration — use our ROI calculator for a detailed breakdown.