How much VRAM does Mistral 7B need?
Mistral 7B needs about 16.1 GB of GPU memory at FP16 and about 5.6 GB at 4-bit. Pick a precision below to see which ComputeBD AI Computer fits and what it costs per hour in taka.
You need about 16.1 GB of GPU memory to run Mistral 7B at FP16 / BF16.
- Model weights 13.5 GB
- Conversation memory (KV cache) 0.5 GB
- Runtime & working space 2.1 GB
NVIDIA L4 is the most affordable AI Computer that fits — ৳147/hour, billed by the minute.
| AI Computer | Fits? | Speed (1 user) | Price |
|---|---|---|---|
| NVIDIA L424 GB | Fits |
≈ 12 tokens/s | ৳147/hr |
| NVIDIA A1024 GB | Fits |
≈ 25 tokens/s | ৳196/hr |
| NVIDIA L40S48 GB | Fits |
≈ 36 tokens/s | ৳343/hr |
| NVIDIA A10080 GB | Fits |
≈ 84 tokens/s | ৳638/hr |
| NVIDIA H10080 GB | Fits |
≈ 139 tokens/s | ৳1,047/hr |
| NVIDIA H200141 GB | Fits |
≈ 199 tokens/s | ৳1,390/hr |
| NVIDIA B200180 GB | Fits |
≈ 331 tokens/s | ৳1,652/hr |
| NVIDIA B300288 GB | Fits |
≈ 331 tokens/s | ৳1,880/hr |
Mistral 7B at each precision
| Precision | GPU memory needed | Cheapest ComputeBD AI Computer |
|---|---|---|
| FP32full precision — rarely needed | 30.7 GB | NVIDIA L40S · ৳343/hr |
| FP16 / BF16standard — original quality | 16.1 GB | NVIDIA L4 · ৳147/hr |
| 8-bitalmost no quality loss | 9.2 GB | NVIDIA L4 · ৳147/hr |
| 4-bit (GGUF / AWQ / GPTQ)small quality loss, ~4× less memory | 5.6 GB | NVIDIA L4 · ৳147/hr |
- Made by
- Mistral AI
- Parameters
- 7.25B
- Max context
- 32,768 tokens
- LoRA / QLoRA fine-tune
- ≈ 17.2 GB / 7.5 GB
- Hugging Face
- mistralai/Mistral-7B-Instruct-v0.3
How the estimate works
- Weights = parameters × bytes per parameter (FP16 = 2 bytes, 8-bit ≈ 1, 4-bit ≈ 0.56).
- KV cache (conversation memory) grows with context length and the number of people at once: 2 × layers × KV-heads × head size × tokens × 2 bytes.
- Runtime: about 1 GB for CUDA plus ~8 % of the weights for working space.
- Fine-tuning: LoRA keeps an FP16 copy of the model and trains ~1 % extra parameters; QLoRA keeps a 4-bit copy; full fine-tuning needs ~16 bytes per parameter. Gradient checkpointing is assumed.
- Speed: generating text for one user is limited by memory bandwidth, so tokens/second ≈ bandwidth × 60 % ÷ model size.
These are good planning estimates, not guarantees: the real figure depends on the software (vLLM, llama.cpp, Transformers…), settings and your data. Leave ~10 % spare.
VRAM needed, model by model
Frequently asked questions
How much VRAM does Mistral 7B need?
About 16.1 GB at FP16 (original quality) and about 5.6 GB at 4-bit, for one user with a 4k context.
Which GPU can run Mistral 7B?
At full quality it fits on an NVIDIA L4 (24 GB), ৳147/hour on ComputeBD.
How much VRAM to fine-tune Mistral 7B?
About 17.2 GB with LoRA and 7.5 GB with QLoRA (sequence length 2,048, batch 1, gradient checkpointing). Estimate the time and cost with the fine-tuning cost estimator.
Try it on a real AI Computer
JupyterLab opens in about a minute. Pay in taka with bKash, billed by the minute — stop whenever you like.