Skip to content

VRAM reference

Qwen3 30B-A3B on an L40S

Yes, you can fine-tune Qwen3 30B-A3B on one L40S (48 GB) with QLoRA. 4-bit inference: yes, 17 GB of weights, room for 128K tokens of context or 64 concurrent 4K-token sequences.

By Will Schulz · Checked 15 November 2026 · Prices as of November 2026

Verdict by method

Fine-tuningMemory at 2K tokensOne L40S?Longest sequence (batch 1)Largest batch at 2KLargest LoRA rank
QLoRA (4-bit base)20 GBYes16K tokens81024
LoRA (16-bit base)64 GBNo———
Full fine-tuning492 GBNo——n/a
InferenceMemory at 8KOne L40S?Longest contextConcurrent 4K sequences
16-bit (BF16/FP16)62 GBNo——
8-bit31 GBYes64K tokens31
4-bit18 GBYes128K tokens64
  • QLoRA: yes, about 20 GB at 2K tokens, batch 1, rank 16; up to 16K-token sequences at batch 1 or batch 8 at 2K, LoRA rank up to 1024.
  • LoRA: no, it needs about 64 GB at 2K tokens, more than the 48 GB available.
  • Full fine-tuning: no, it needs about 492 GB at 2K tokens, more than the 48 GB available.

Where the memory goes

At 2,048 tokens per sequence, batch 1, LoRA rank 16, gradient checkpointing on
MethodWeightsTraining stateActivationsTotalOn one L40S80 GB GPUs needed
QLoRA (4-bit base)17 GB0.2 GB3.1 GB20 GBFits1
LoRA (16-bit base)61 GB0.2 GB3.1 GB64 GBDoesn't fit1
Full fine-tuning61 GB427 GB3.1 GB492 GBDoesn't fit7
One sequence, fp16 KV cache
PrecisionWeightsKV cache at 8KTotalAt 128K contextOn one L40S
16-bit (BF16/FP16)61 GB0.8 GB62 GB74 GBDoesn't fit
8-bit31 GB0.8 GB31 GB43 GBFits
4-bit17 GB0.8 GB18 GB30 GBFits

The cheapest way to run it

Renting an L40S costs $1.09 an hour on Runpod Secure Cloud ($0.79 on Runpod Community Cloud). QLoRA on it processes about 8,556 tokens a second by our estimate, so 10 million training tokens take 19 minutes and cost about $0.35.

Job (default settings)MemoryCheapest rental that fitsPer hour
QLoRA fine-tuning20 GBRTX 3090$0.50
LoRA fine-tuning64 GBA100 80GB$2.19
4-bit inference18 GBRTX 3090$0.50
16-bit inference62 GBA100 80GB$2.19

On-demand rates seen on 15 November 2026: Runpod Secure Cloud, Lambda for the A100 40GB. Consumer cards count as single GPUs only; data-centre GPUs up to eight per node. Kojen's GPU training can launch the L40S on Runpod, Vast.ai or Lambda, or pick the cheapest GPU in stock for you.

Try other settings

JobMemoryL40S (48 GB)Smallest single GPU
QLoRA (4-bit base)2K tokens × batch 1, rank 1620 GBFitsRTX 3090 (24 GB)
LoRA (16-bit base)2K tokens × batch 1, rank 1664 GB2 GPUsA100 80GB (80 GB)
Full fine-tuning2K tokens × batch 1492 GB12 GPUsNone: needs several
Inference, 16-bit (BF16/FP16)8K-token context × 1 sequence62 GB2 GPUsA100 80GB (80 GB)
Inference, 8-bit8K-token context × 1 sequence31 GBFitsA100 40GB (40 GB)
Inference, 4-bit8K-token context × 1 sequence18 GBFitsRTX 3090 (24 GB)

A job fits when the estimate is within 90% of the card's memory. Training assumes gradient checkpointing and FlashAttention; multi-GPU counts assume the memory splits evenly (FSDP, ZeRO-3 or tensor parallelism).

How we estimate it

Every figure comes from the model's published architecture (layers, hidden size, attention and KV heads, vocabulary) rather than a rule of thumb, using the same functions as the calculator above. Sizes are in GB (10⁹ bytes).

  • Full fine-tuning: 16 bytes per parameter for mixed-precision Adam (bf16 weights and gradients, fp32 master weights and two moments), as in the ZeRO paper.
  • LoRA: frozen bf16 weights (2 bytes per parameter) plus 16 bytes per adapter parameter, rank 16 on every linear layer. For mixture-of-experts models, adapters go on attention only and the experts stay frozen.
  • QLoRA: NF4 weights with double quantisation (4.127 bits per parameter); embeddings and the output head stay in bf16, as bitsandbytes leaves them.
  • Activations: gradient checkpointing and FlashAttention. Each layer's bf16 input is kept, one layer is recomputed at a time, and the output logits take 8 bytes per vocabulary entry per token (fp32 values and gradients). Chunked cross-entropy, as in Unsloth or Liger kernels, cuts that last term.
  • Inference: 2, 1 or about 0.56 bytes per weight (4-bit with group-wise scales), plus an fp16 KV cache of 2 × layers × KV heads × head dim × 2 bytes per token; sliding-window layers only cache their window.
  • Fits means the estimate is within 90% of the card's listed memory; the rest covers the CUDA context, kernels and fragmentation. Multi-GPU counts assume memory splits evenly.

Speed estimates use L40S's dense BF16 peak of 362 TFLOPS, about 4 FLOPs per active parameter per token for LoRA and 40% utilisation; see fine-tuning and training costs for the method.

Frequently asked questions

Can I fine-tune Qwen3 30B-A3B on an L40S?
QLoRA: yes, about 20 GB at 2K tokens, batch 1, rank 16; up to 16K-token sequences at batch 1 or batch 8 at 2K, LoRA rank up to 1024. LoRA: no, it needs about 64 GB at 2K tokens, more than the 48 GB available. Full fine-tuning: no, it needs about 492 GB at 2K tokens, more than the 48 GB available.
Can I run Qwen3 30B-A3B on an L40S?
16-bit (BF16/FP16) inference: no, the weights alone are 61 GB. 8-bit inference: yes, 31 GB of weights, room for 64K tokens of context or 31 concurrent 4K-token sequences. 4-bit inference: yes, 17 GB of weights, room for 128K tokens of context or 64 concurrent 4K-token sequences.
What does fine-tuning Qwen3 30B-A3B on an L40S cost?
With QLoRA at our planning rate of 40% utilisation, about 19 minutes for 10 million training tokens, around $0.35 at $1.09 an hour (Runpod Secure Cloud, November 2026).

Sources

  1. [1]Hugging Face. Qwen/Qwen3-30B-A3B config.json
  2. [2]Alibaba Qwen. Qwen/Qwen3-30B-A3B model card
  3. [3]Alibaba Qwen. Qwen3 Technical Report (arXiv:2505.09388)
  4. [4]NVIDIA. L40S specifications
  5. [5]Runpod. GPU pricing (updated 12 November 2026)
  6. [6]Lambda. GPU cloud pricing
  7. [7]Dettmers et al., 2023. QLoRA: Efficient Finetuning of Quantized LLMs (arXiv:2305.14314)
  8. [8]Hu et al., 2021. LoRA: Low-Rank Adaptation of Large Language Models (arXiv:2106.09685)
  9. [9]Rajbhandari et al., 2019. ZeRO: Memory Optimizations Toward Training Trillion Parameter Models (arXiv:1910.02054)
  10. [10]Korthikanti et al., 2022. Reducing Activation Recomputation in Large Transformer Models (arXiv:2205.05198)

Measure before you rent the GPU.

Prompt tuning and model search in Kojen run on hosted models against your own test cases, so you know whether fine-tuning Qwen3 30B-A3B is worth the GPU time.

Start training free