How It Works Under the Hood
Model weights take parameters × bits per weight ÷ 8 bytes. The KV cache grows with context: 2 × layers × KV heads × head dimension × tokens × batch × bytes per value. The tool adds a runtime overhead of about 10% of the weights plus 0.5 GiB for CUDA context and activations, then compares the total with the VRAM you have.
Step-by-Step Instructions
Pick a preset model or choose custom and enter parameters, layers, KV heads and head dimension.
Choose the quantization (FP16, Q8, Q6, Q5, Q4 or Q3 style) and the KV cache precision.
Set the context length and batch size you plan to use.
Enter the VRAM per GPU and the number of GPUs to see whether the model fits and how much is left.
Key Features & Performance
- Presets for Llama 3.1, Mistral 7B, Qwen2.5 and Gemma 2 with their real layer and KV head counts
- Separate lines for weights, KV cache and overhead so you see what to cut first
- Fit check against common GPU sizes and multi-GPU setups
- Runs locally; nothing you type is sent anywhere
Methodology, Formulas & Technical Standards
Effective bits per weight for GGUF-style quantizations are averages (Q4_K_M is about 4.85 bits, Q8_0 about 8.5). Real files differ slightly by architecture.
Common Real-World Use Cases
- Deciding whether a 24 GB card can run a 32B model at 4-bit with a long context.
- Sizing a multi-GPU server before renting hardware.
- Seeing how much memory a longer context window costs.
Worked Examples & Edge Cases
Llama 3.1 8B KV cache
Llama 3.1 8B, 8,192 tokens, batch 1, FP16 cache
1.00 GiB of KV cache
128 KiB per token: 2 × 32 layers × 8 KV heads × 128 dims × 2 bytes.
Weights of a 70B model
70B parameters at 16 bits
140 GB of weights
Far beyond one consumer GPU, which is why 70B is usually run at 4-bit or split across cards.
Limitations, Scope & Accuracy Notes
- • This is an estimate. Inference engines differ in overhead, and activation memory depends on the backend and batch shape.
- • It does not model offloading to system RAM, tensor-parallel duplication or training memory (gradients and optimizer state).
Local Browser Processing & Privacy
Traditional web converters upload your documents, snippets, or media files to remote servers. LLM VRAM Requirement Calculator runs inside your browser tab:
Your input text and files are processed locally without uploading payloads to a processing server.
Microdeck does not log or store the files or inputs you paste into tool workspaces (normal static hosting/CDN requests still load page assets).
Working outputs stay in your browser tab and are released when you reset the tool or close the tab.
Frequently Asked Questions
Why does the KV cache matter?
It grows linearly with context length and batch size. At very long contexts it can rival the weights, so a model that fits at 4K tokens may not fit at 128K.
Does a 4-bit model lose quality?
Usually a little. Q4_K_M and Q5 variants are widely used because the quality loss is small compared with the memory saved, but it depends on the model and task.
Can I run a model that is larger than my VRAM?
Some runtimes offload layers to system RAM, which works but is much slower. This tool only checks whether everything fits on the GPU.
Explore Related Micro-Tools
LLM Tokenizer & Cost Estimator
Estimate token counts & reference API / VRAM costs across modern LLMs
Storage SI/IEC & Bandwidth Transfer Calculator
Convert Bytes, KB/KiB, MB/MiB, GB/GiB, TB/TiB & Bits simultaneously and estimate file transfer times with TCP overhead
Structured LLM Output & Schema Generator
Convert JSON payloads into Zod, Pydantic & JSON Schema