Skip to content
Updated minutes ago
Alibaba

Qwen 2.5 Coder 3B

Alibaba · dense · 3.1B parameters · 32,768 context

Quality
50.0

Parameters

3.1B

Context Window

32K tokens

Architecture

Dense

Best GPU

RTX 4070 Ti

Quality Score

50/100

Intelligence Brief

Qwen 2.5 Coder 3B is a 3.1B parameter DENSE model from Alibaba, featuring Grouped Query Attention (GQA) with 36 layers and 2,048 hidden dimensions. With a 32,768 token context window, it supports structured output, code, math. On standardized benchmarks, it achieves MMLU 55, HumanEval 40, GSM8K 55. For self-hosted inference, RTX 4070 Ti delivers optimal throughput at $237/month.

Architecture Details

TypeDENSE
Total Parameters3.1B
Active Parameters3.1B
Layers36
Hidden Dimension2,048
Attention Heads16
KV Heads2
Head Dimension128
Vocab Size152,064

Memory Requirements

BF16 Weights

6.2 GB

FP8 Weights

3.1 GB

INT4 Weights

1.6 GB

KV-Cache per Token18432 bytes
Activation Estimate0.40 GB

GPU Compatibility Matrix

Qwen 2.5 Coder 3B is compatible with 100% 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

RTX 4070 Tioptimal

BF16 · 1 GPU · vllm

100/100

score

Throughput

438.9 tok/s

Latency (ITL)

2.3ms

Est. TTFT

0ms

Cost/Month

$237

Cost/M Tokens

$0.21

Use this config →
RTX 3080optimal

BF16 · 1 GPU · vllm

100/100

score

Throughput

661.9 tok/s

Latency (ITL)

1.5ms

Est. TTFT

0ms

Cost/Month

$133

Cost/M Tokens

$0.08

Use this config →
RTX 4070 Superoptimal

BF16 · 1 GPU · vllm

100/100

score

Throughput

438.9 tok/s

Latency (ITL)

2.3ms

Est. TTFT

0ms

Cost/Month

$209

Cost/M Tokens

$0.18

Use this config →

Deployment Options

API

API Deployment

No API pricing available

Self-Hosted

Single GPU

RTX 4070 Ti

$237/mo

Min VRAM: 3 GB

Scale

Multi-GPU

RTX 4070 Ti

438.9 tok/s

Best available config

API Pricing Comparison

No API pricing data available for this model.

Performance Estimates

Throughput by GPU

RTX 4070 Ti
438.9 tok/s
RTX 3080
661.9 tok/s
RTX 4070 Super
438.9 tok/s

VRAM Breakdown (RTX 4070 Ti, BF16)

Weights
Act
Weights 6.2 GBKV-Cache 0.6 GBActivations 3.2 GBOverhead 0.5 GB

Precision Impact

bf16

6.2 GB

weights/GPU

~438.9 tok/s

fp8

3.1 GB

weights/GPU

int4

1.6 GB

weights/GPU

Quality Benchmarks

Bottom 25%
7th percentile across all models
MMLU
55.0
Bottom 25% (14th pctile)
HumanEval
40.0
Below Average (25th pctile)
GSM8K
55.0
Bottom 25% (19th pctile)
MT-Bench
68.0
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 2.5 Coder 3B

Similar Models

Lower quality, Smaller modelCompare →
Larger context, Larger modelCompare →

Qwen 2.5 3B

3.09B params · dense

Quality: 58

from $0.10/M

Higher qualityCompare →

StarCoder2 3B

3.03B params · dense

Quality: 29

from $0.10/M

Lower qualityCompare →
Smaller contextCompare →

Frequently Asked Questions

How much VRAM does Qwen 2.5 Coder 3B need for inference?

Qwen 2.5 Coder 3B requires approximately 6.2 GB of VRAM at BF16 precision, 3.1 GB at FP8, or 1.6 GB at INT4 quantization. Additional VRAM is needed for KV-cache (18432 bytes per token) and activations (~0.40 GB).

What is the best GPU for Qwen 2.5 Coder 3B?

The top recommended GPU for Qwen 2.5 Coder 3B is the RTX 4070 Ti using BF16 precision. It achieves approximately 438.9 tokens/sec at an estimated cost of $237/month ($0.21/M tokens). Score: 100/100.

How much does Qwen 2.5 Coder 3B inference cost?

Qwen 2.5 Coder 3B inference costs vary by provider and GPU setup. Use our calculator for detailed cost estimates across all providers.