▶ Fine-tune Calculator▋

Estimate the GPU memory and cost to fine-tune and serve an LLM — full vs LoRA vs QLoRA, multi-GPU splits, serving concurrency, cloud vs local, and download size.

• 01Fine-tuning VRAM Calculator — Full vs LoRA vs QLoRAEstimate the GPU memory to fine-tune an LLM, broken down into base weights, gradients, optimizer state and activations.training
• 02Multi-GPU LLM Split CalculatorHow many GPUs you need to hold a model, and the VRAM per card at tensor-parallel 2, 4 and 8.multi-gpu
• 03LLM Serving Concurrency CalculatorHow many concurrent requests a GPU can serve — from the VRAM left after the weights, divided by the KV cache per request.serving
• 04Cloud GPU vs Local — LLM Cost CalculatorThe smallest cloud GPU tier that fits your model, its dollars per million tokens, and how that compares to an API.cost
• 05LLM Download Size & Disk CalculatorHow much disk a model takes at a given quantization, how many files it splits into, and how long the download takes.storage

Sibling of vram-calc.com, which covers plain inference sizing. Every tool takes model / GPU presets or custom numbers; presets carry a verified date.