Mistral 7B VRAM requirements
Fine-tuning Mistral 7B takes about 6.3 GB of GPU memory with QLoRA, 17 GB with LoRA on 16-bit weights and 117 GB for full fine-tuning, at 2,048 tokens per sequence. Running it takes 16 GB in 16-bit or 5.2 GB in 4-bit with a 8K-token context.
Fine-tuning memory
| Method | Weights | Training state | Activations | Total | Smallest single GPU | 80 GB GPUs needed |
|---|---|---|---|---|---|---|
| QLoRA (4-bit base) | 4.1 GB | 0.7 GB | 1.5 GB | 6.3 GB | RTX 3090 (24 GB) | 1 |
| LoRA (16-bit base) | 14 GB | 0.7 GB | 1.5 GB | 17 GB | RTX 3090 (24 GB) | 1 |
| Full fine-tuning | 14 GB | 101 GB | 1.5 GB | 117 GB | H200 (141 GB) | 2 |
On a 24 GB card such as the RTX 4090: QLoRA: yes, about 6.3 GB at 2K tokens, batch 1, rank 16; up to 16K-token sequences at batch 1 or batch 11 at 2K, LoRA rank up to 256.
Inference memory
| Precision | Weights | KV cache at 8K | Total | At 32K context | Smallest single GPU |
|---|---|---|---|---|---|
| 16-bit (BF16/FP16) | 14 GB | 1.1 GB | 16 GB | 19 GB | RTX 3090 (24 GB) |
| 8-bit | 7.2 GB | 1.1 GB | 8.3 GB | 12 GB | RTX 3090 (24 GB) |
| 4-bit | 4.1 GB | 1.1 GB | 5.2 GB | 8.4 GB | RTX 3090 (24 GB) |
Which GPUs fit Mistral 7B
One GPU unless a count is shown. Follow a GPU's name for its maximum sequence length, batch size, LoRA rank and rental cost.
| GPU | QLoRA | LoRA | Full fine-tuning | Run, fp16 | Run, int8 | Run, int4 |
|---|---|---|---|---|---|---|
| RTX 3090 24 GB | Yes | Yes | 6× | Yes | Yes | Yes |
| RTX 4090 24 GB | Yes | Yes | 6× | Yes | Yes | Yes |
| RTX 5090 32 GB | Yes | Yes | 5× | Yes | Yes | Yes |
| L40S 48 GB | Yes | Yes | 3× | Yes | Yes | Yes |
| A100 40GB 40 GB | Yes | Yes | 4× | Yes | Yes | Yes |
| A100 80GB 80 GB | Yes | Yes | 2× | Yes | Yes | Yes |
| H100 SXM 80 GB | Yes | Yes | 2× | Yes | Yes | Yes |
| H200 141 GB | Yes | Yes | Yes | Yes | Yes | Yes |
| B200 180 GB | Yes | Yes | Yes | Yes | Yes | Yes |
What's specific to Mistral 7B
- v0.3 drops the sliding window of the original release and uses full attention over 32K tokens.
- A 32K-token vocabulary keeps training logits small, so it trains at longer sequence lengths than Llama 3.1 8B on the same card.
| Architecture | Mistral 7B |
|---|---|
| Parameters | 7.25B |
| Layers | 32 |
| Hidden size | 4,096 |
| MLP width | 14,336 |
| Attention heads / KV heads | 32 / 8 (head dim 128) |
| Vocabulary | 32,768 tokens |
| Context | 32K tokens |
| License | Apache 2.0 |
| Released | May 2024 (v0.3) |
Try other settings
Pre-filled for Mistral 7B. Change the sequence length, batch, LoRA rank, context or GPU.
| Job | Memory | RTX 4090 (24 GB) | Smallest single GPU |
|---|---|---|---|
| QLoRA (4-bit base)2K tokens × batch 1, rank 16 | 6.3 GB | Fits | RTX 3090 (24 GB) |
| LoRA (16-bit base)2K tokens × batch 1, rank 16 | 17 GB | Fits | RTX 3090 (24 GB) |
| Full fine-tuning2K tokens × batch 1 | 117 GB | 6 GPUs | H200 (141 GB) |
| Inference, 16-bit (BF16/FP16)8K-token context × 1 sequence | 16 GB | Fits | RTX 3090 (24 GB) |
| Inference, 8-bit8K-token context × 1 sequence | 8.3 GB | Fits | RTX 3090 (24 GB) |
| Inference, 4-bit8K-token context × 1 sequence | 5.2 GB | Fits | RTX 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.
Frequently asked questions
- Can I fine-tune Mistral 7B on a 24 GB GPU?
- On an RTX 3090 or 4090 (24 GB): QLoRA: yes, about 6.3 GB at 2K tokens, batch 1, rank 16; up to 16K-token sequences at batch 1 or batch 11 at 2K, LoRA rank up to 256. LoRA: yes, about 17 GB at 2K tokens, batch 1, rank 16; up to 8K-token sequences at batch 1 or batch 4 at 2K, LoRA rank up to 128. Estimates use gradient checkpointing and leave 10% of memory free.
- How much VRAM do you need to run Mistral 7B?
- About 16 GB in 16-bit and 5.2 GB in 4-bit, at 8K tokens of context. The smallest single GPU here for 4-bit is the RTX 3090 (24 GB); for 16-bit, the RTX 3090 (24 GB).
- How many GPUs does full fine-tuning Mistral 7B take?
- About 117 GB with mixed-precision Adam at 2K tokens, so at least 2 × 80 GB H100s with the training state sharded (FSDP or ZeRO-3). LoRA needs 17 GB and QLoRA 6.3 GB.
- How long a context can Mistral 7B handle on an 80 GB GPU?
- Its KV cache costs 4.3 GB at the full 32K tokens (fp16 cache). On one 80 GB GPU: 32K tokens with 16-bit weights, 32K tokens with 4-bit weights.
Related
- Mistral 7B on RTX 309024 GB · max settings and cost
- Mistral 7B on RTX 409024 GB · max settings and cost
- Mistral 7B on RTX 509032 GB · max settings and cost
- What fine-tuning Mistral 7B costsGPU-hours and dollars per million tokens
- Mixtral 8x7B VRAMMistral AI
Terms used on this page: LoRA, QLoRA, KV cache, gradient checkpointing, quantization. For any model size, use the fine-tuning VRAM calculator; for the method in depth, the Academy lesson Memory first: what hardware can train what.
Sources
- Hugging Face. mistralai/Mistral-7B-Instruct-v0.3 config.json
- Mistral AI. mistralai/Mistral-7B-Instruct-v0.3 model card
- Mistral AI. Mistral 7B (arXiv:2310.06825)
- Dettmers et al., 2023. QLoRA: Efficient Finetuning of Quantized LLMs (arXiv:2305.14314)
- Hu et al., 2021. LoRA: Low-Rank Adaptation of Large Language Models (arXiv:2106.09685)
- Rajbhandari et al., 2019. ZeRO: Memory Optimizations Toward Training Trillion Parameter Models (arXiv:1910.02054)
- Korthikanti et al., 2022. Reducing Activation Recomputation in Large Transformer Models (arXiv:2205.05198)
- Chen et al., 2016. Training Deep Nets with Sublinear Memory Cost (arXiv:1604.06174)
Architecture read from config.json and parameter count from the Hugging Face safetensors metadata on 15 November 2026.
Find out if you need to fine-tune at all.
Kojen's prompt tuning and model search run for real on your own test cases, on hosted models. Measure the cheaper levers before you rent a GPU.
Start training free