Skip to content

Best GPU for Qwen3 Coder 30B A3B: VRAM Requirements

Qwen3 Coder 30B A3B needs about 68 GB of GPU memory in BF16 and 17.7 GB at 4 bits. See which GPUs fit it for inference and fine-tuning.

Qwen3 Coder 30B A3B needs about 68 GB of VRAM in FP16, 34.4 GB in FP8 and 17.7 GB at 4 bits for inference with an 8,192-token context. In FP16 it fits on one NVIDIA A100 PCIe 80GB; a 4-bit build fits on one NVIDIA GeForce RTX 4090 24GB.

Qwen3 Coder 30B A3B is a mixture-of-experts language model with 30.5B total and 3.3B active parameters per token from Alibaba Cloud (Qwen team), released in July 2025, with a 262,144-token context window and published under the Apache 2.0. Its weights alone take 61 GB in BF16 and 15.3 GB at 4 bits per parameter; the tables below add the KV cache and runtime memory, then list the GPUs that hold each setup.

Recommended GPUs for Qwen3 Coder 30B A3B

Qwen3 Coder 30B A3B VRAM requirements by precision

One request with an 8,192-token context, KV cache in BF16, plus a 10 % margin for activations and runtime buffers.

PrecisionWeightsKV cacheTotal for inference
BF16 / FP1661 GB0.8 GB68 GB
FP830.5 GB0.8 GB34.4 GB
INT415.3 GB0.8 GB17.7 GB

Qwen3 Coder 30B A3B VRAM by task

Fine-tuning rows assume a short sequence, batch 1 and gradient checkpointing; longer sequences add activation memory.

TaskGPU memory
Inference at the full 262,144-token context (FP8)61.9 GB
QLoRA (4-bit base)19.8 GB
LoRA (BF16 base)79.3 GB
Full fine-tuning (BF16, AdamW)536.8 GB

GPUs that fit Qwen3 Coder 30B A3B

Number of cards whose combined memory holds each setup, assuming the model is split across them with tensor or pipeline parallelism. A dash means more than eight cards.

GPUMemoryBF16 inference (68 GB)FP8 inference (34.4 GB)INT4 inference (17.7 GB)QLoRA (19.8 GB)LoRA (79.3 GB)
NVIDIA GeForce RTX 5080 16GB16 GB8 GPUs4 GPUs2 GPUs2 GPUs8 GPUs
NVIDIA GeForce RTX 4090 24GB24 GB4 GPUs2 GPUs1 GPU1 GPU4 GPUs
NVIDIA L4 24GB24 GB4 GPUs2 GPUs1 GPU1 GPU4 GPUs
NVIDIA GeForce RTX 5090 32GB32 GB4 GPUs2 GPUs1 GPU1 GPU4 GPUs
NVIDIA Tesla V100 32GB32 GB4 GPUs2 GPUs1 GPU1 GPU4 GPUs
NVIDIA A100 PCIe 40GB40 GB2 GPUs1 GPU1 GPU1 GPU2 GPUs
NVIDIA L40S 48GB48 GB2 GPUs1 GPU1 GPU1 GPU2 GPUs
NVIDIA Quadro RTX 8000 48GB48 GB2 GPUs1 GPU1 GPU1 GPU2 GPUs
NVIDIA RTX 6000 Ada 48GB48 GB2 GPUs1 GPU1 GPU1 GPU2 GPUs
NVIDIA RTX A6000 48GB48 GB2 GPUs1 GPU1 GPU1 GPU2 GPUs
NVIDIA A100 PCIe 80GB80 GB1 GPU1 GPU1 GPU1 GPU1 GPU
NVIDIA A100 SXM 80GB80 GB1 GPU1 GPU1 GPU1 GPU1 GPU
NVIDIA H100 PCIe 80GB80 GB1 GPU1 GPU1 GPU1 GPU1 GPU
NVIDIA H100 SXM 80GB80 GB1 GPU1 GPU1 GPU1 GPU1 GPU
NVIDIA RTX PRO 6000 Blackwell 96GB96 GB1 GPU1 GPU1 GPU1 GPU1 GPU
NVIDIA H200 141GB141 GB1 GPU1 GPU1 GPU1 GPU1 GPU
NVIDIA B200 180GB180 GB1 GPU1 GPU1 GPU1 GPU1 GPU
NVIDIA B300 288GB288 GB1 GPU1 GPU1 GPU1 GPU1 GPU
NVIDIA GB300 NVL288 GB1 GPU1 GPU1 GPU1 GPU1 GPU

What is Qwen3 Coder 30B A3B?

Qwen3 Coder 30B A3B Instruct is the compact model of the Qwen3 Coder series: a mixture-of-experts model with 30.5 billion parameters, 3.3 billion of them active per token, built for agentic coding, repository-scale code understanding and tool calling.

Developer
Alibaba Cloud (Qwen team)
Released
July 2025
Parameters
30.5B total, 3.3B active per token
Context window
262,144 tokens
License
Apache 2.0

The native window is 262,144 tokens; the model card extends it to about 1 million tokens with YaRN scaling.

Source: Qwen3 Coder 30B A3B Instruct model card (as of November 17, 2026).

Frequently asked questions

How much VRAM does Qwen3 Coder 30B A3B need?
For inference with an 8,192-token context and one request, Qwen3 Coder 30B A3B needs about 68 GB in BF16, 34.4 GB in FP8, 17.7 GB in INT4, including a 10 % runtime margin. Real usage depends on the inference engine and its settings.
Can Qwen3 Coder 30B A3B run on a single GPU?
Yes. In BF16 it needs about 68 GB, which fits one NVIDIA A100 PCIe 80GB (80 GB).
How much GPU memory does fine-tuning Qwen3 Coder 30B A3B take?
QLoRA (4-bit base): about 19.8 GB; LoRA (BF16 base): about 79.3 GB; Full fine-tuning (BF16, AdamW): about 536.8 GB. Longer training sequences and larger batches add activation memory on top.
What is the best GPU for Qwen3 Coder 30B A3B?
Best for inference: 1 GPU × NVIDIA A100 PCIe 40GB (FP8 or INT8 weights, 34.4 GB at 8,192 tokens). Best for fine-tuning: 1 GPU × NVIDIA H200 141GB (LoRA on BF16 weights, 79.3 GB). Budget pick: 1 GPU × NVIDIA GeForce RTX 4090 24GB (Smallest card class that holds a 4-bit build, 17.7 GB). Check live offers before booking, as availability changes.
Can Qwen3 Coder 30B A3B run on an RTX 4090?
Yes. In 4-bit it needs about 17.7 GB, which fits one RTX 4090 (24 GB). FP16 needs about 68 GB, 4 of these cards.
Can Qwen3 Coder 30B A3B run on an RTX 5090?
Yes. In 4-bit it needs about 17.7 GB, which fits one RTX 5090 (32 GB). FP16 needs about 68 GB, 4 of these cards.

Estimates use the formulas of the LLM VRAM calculator (decimal GB). Real usage depends on the inference engine, quantization format and memory fragmentation: test on the target GPU before a long rental.

Explore the live market

Every offer on the live market comes from a reviewed supplier and is sold by EverXCloud Inc.