Tools
Everything to size, cost and choose your AI hardware.
One engine, cited and parity-tested, behind every calculator. Start with a question — will it run, what will it cost, which card wins — and get a straight answer.
Calculators
AI Hardware Advisor
The one-page answer: describe your machine and budget in a sentence and get what runs today, the best picks for your goal, what money unlocks, and whether to rent.
Open →WizardGPU Recommendation Wizard
No specs knowledge needed: answer six quick questions — budget, use case, local, power, brand, country — and get your card with alternatives, pros & cons and an upgrade path.
Open →Reality check“Why Not?” GPU Check
Every card has a ceiling. Pick any GPU and see exactly where it runs out of room — 70B models, 128K context, fine-tuning — with the cheapest better option that clears each limit.
Open → BuilderAI Hardware Builder
Set a budget and pick your workloads — get a full local-AI rig (GPU, CPU, PSU, RAM, SSD) with tokens/sec, context, running cost and an upgrade path.
Open → FitCan I Run It?
Pick a GPU and a model — get Yes / No / “only at Q4”, with the exact VRAM breakdown and decode speed.
Open → VRAMVRAM Calculator
Exact inference memory for any open model at any quant — weights, KV cache and overhead.
Open → TrainingFine-Tuning Calculator
Full, LoRA and QLoRA memory for training runs — optimizer, gradients and activations included.
Open → CostGPU Cost Calculator
True cost of ownership: sticker, electricity and amortised monthly — with an own-vs-rent verdict.
Open → DecideBuy vs Rent
Pick a model and see whether buying a local rig or renting cloud GPUs wins — exact hardware, cloud rate and the break-even month.
Open → ROITraining ROI
Break-even for a training workload: hours, power and cloud cost per run, and when owning pays off.
Open → RigsMulti-GPU Planner
2× 4090 or 4× H100 — pooled VRAM, real tensor-parallel speed, wall power and full rig cost.
Open → BuildPC Compatibility
Enter CPU, RAM, motherboard and PSU — see which GPUs actually fit, with a real power budget.
Open → OllamaOllama VRAM
Memory for local Ollama models, tuned to how Ollama loads and quantizes them.
Open →Decision tools
AI Model Finder
Tell it the job — coding, chat, reasoning — and it ranks the famous open models on speed, VRAM, accuracy and license.
Open → CompareGPU Compare
Up to three cards side by side — fit, decode speed, perf-per-dollar and buy-vs-cloud break-even.
Open → RankThe Bitpute Index
Proprietary GPU scores — VRAM efficiency, tokens per dollar, perf per watt and a Local Inference Score.
Open →One engine behind all of it
Numbers you can check.
Every tool runs on the same cited, parity-tested memory engine. See exactly how the math works.