Skip to content
Updated minutes ago
Alibaba

Qwen 3 30B-A3B

Alibaba · moe · 30.5B parameters · 131,072 context

Quality
70.0

Parameters

30.5B

Context Window

128K tokens

Architecture

MoE

Best GPU

H200 SXM

Quality Score

70/100

Intelligence Brief

Qwen 3 30B-A3B is a 30.5B parameter Mixture-of-Experts (128 experts, 8 active) model from Alibaba, featuring Grouped Query Attention (GQA) with 48 layers and 2,048 hidden dimensions. With a 131,072 token context window, it supports tools, structured output, code, math, multilingual, reasoning. On standardized benchmarks, it achieves MMLU 75, HumanEval 48, GSM8K 80. For self-hosted inference, H200 SXM delivers optimal throughput at $2553/month.

Architecture Details

TypeMOE
Total Parameters30.5B
Active Parameters3.3B
Layers48
Hidden Dimension2,048
Attention Heads32
KV Heads4
Head Dimension128
Vocab Size151,936
Total Experts128
Active Experts8

Memory Requirements

BF16 Weights

61.0 GB

FP8 Weights

30.5 GB

INT4 Weights

15.3 GB

KV-Cache per Token24576 bytes
Activation Estimate0.50 GB

GPU Compatibility Matrix

Qwen 3 30B-A3B is compatible with 62% 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

H200 SXMoptimal

FP8 · 1 GPU · tensorrt-llm

100/100

score

Throughput

1.1K tok/s

Latency (ITL)

1.0ms

Est. TTFT

0ms

Cost/Month

$2553

Cost/M Tokens

$0.93

Use this config →
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

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 →

Deployment Options

API

API Deployment

No API pricing available

Self-Hosted

Single GPU

H200 SXM

$2553/mo

Min VRAM: 31 GB

Scale

Multi-GPU

RTX A6000 x2

878.7 tok/s

TP· $930/mo

API Pricing Comparison

No API pricing data available for this model.

Performance Estimates

Throughput by GPU

H200 SXM
1.1K tok/s
H100 SXM
1.1K tok/s
H100 PCIe
1.1K tok/s

VRAM Breakdown (H200 SXM, FP8)

Weights
Weights 30.5 GBKV-Cache 0.8 GBActivations 4.0 GBOverhead 1.5 GB

Precision Impact

bf16

61.0 GB

weights/GPU

fp8

30.5 GB

weights/GPU

~1.1K tok/s

int4

15.3 GB

weights/GPU

Quality Benchmarks

Average
73th percentile across all models
MMLU
75.0
Below Average (46th pctile)
HumanEval
48.0
Below Average (43th pctile)
GSM8K
80.0
Below Average (47th pctile)
MT-Bench
78.0
Bottom 25% (0th pctile)

Capabilities

Features

Tool Use Vision Code Math Reasoning Multilingual Structured Output

Supported Frameworks

vllmsglangtgiollama

Supported Precisions

BF16 (default)FP8INT4

Where to Deploy Qwen 3 30B-A3B

Similar Models

Qwen 3 32B

32.8B params · dense

Quality: 74

from $0.80/M

Similar specsCompare →
Smaller context, Lower qualityCompare →
Smaller context, Lower qualityCompare →

Gemma 4 31B-IT

31B params · dense

Quality: 77

from $0.30/M

Smaller context, Higher qualityCompare →

Qwen 2.5 32B

32.5B params · dense

Quality: 73

from $0.80/M

Similar specsCompare →

Frequently Asked Questions

How much VRAM does Qwen 3 30B-A3B need for inference?

Qwen 3 30B-A3B requires approximately 61.0 GB of VRAM at BF16 precision, 30.5 GB at FP8, or 15.3 GB at INT4 quantization. Additional VRAM is needed for KV-cache (24576 bytes per token) and activations (~0.50 GB).

What is the best GPU for Qwen 3 30B-A3B?

The top recommended GPU for Qwen 3 30B-A3B is the H200 SXM using FP8 precision. It achieves approximately 1.1K tokens/sec at an estimated cost of $2553/month ($0.93/M tokens). Score: 100/100.

How much does Qwen 3 30B-A3B inference cost?

Qwen 3 30B-A3B inference costs vary by provider and GPU setup. Use our calculator for detailed cost estimates across all providers.