▚ VRAM Calculator▋

Estimate the GPU VRAM to run any LLM — by model, quantization and context length. Plus a will-it-fit checker, quant comparison, KV-cache and tokens/sec tools.

• 01LLM VRAM CalculatorEstimate the GPU memory to run any large language model — from the parameter count, quantization format and context length.memory
• 02Will It Fit? — GPU vs LLM CheckerPick your GPU and a model, get a straight fits / fits-with-offload / won't-fit answer, plus the largest quant that fits fully.fit check
• 03LLM Quantization ComparisonOne model, every quant format side by side — file size, total VRAM, whether it fits your card, and the rough quality cost.quantization
• 04KV Cache & Max Context CalculatorHow much memory the KV cache costs per token for a given model, and the longest context that fits in your spare VRAM.context
• 05LLM Tokens per Second EstimatorEstimate decode speed and time-to-first-token for a model on a given GPU — from memory bandwidth, not TFLOPs.speed
• 06Mac Unified Memory LLM CalculatorCan your Apple Silicon Mac run a given model? Accounts for the OS reserve and the GPU wired-memory limit.apple silicon

Model architectures are pulled straight from Hugging Face configs and refreshed every few days; the KV-cache math is computed per attention type — standard (GQA/MHA), MLA latent cache, sliding-window and hybrid attention/SSM. Every tool also takes fully custom numbers. Training, multi-GPU, serving and cost sizing live on the sibling sitefinetune-calc.com.