Just in

Local AI · Topic hub

GPUs & Hardware

VRAM maths, RTX buying guides and what actually predicts local performance.

31 articles across guides, comparisons

A single RTX 5090 lit from below on a dark test bench, its heatsink fins glowing faintly while a monitor behind it shows a stalled token-generation graph against a soaring prefill curve.Guides

RTX 5090 AI TOPS Explained: What 3352 Actually Measures

NVIDIA's 3352 AI TOPS is sparse FP4, not the precision your local model runs. The full ladder, the 16x gap, and when TOPS is the wrong spec to buy on.

15 Aug11 min read
A close-up cinematic shot of an open PC case lit by amber GPU LEDs, a single triple-slot graphics card seated in the top PCIe lane with an empty second slot beside it, dust visible on the shroud, monitor glow reflecting off the side panel.Guides

VRAM for Training vs Inference: The 8x Byte Ledger

Training costs 16 bytes per parameter; inference costs 2. Here is the full byte ledger, why the multiplier is a range, and which GPU it rules out.

14 Aug8 min read
Wide shot down a data centre aisle of glowing GPU racks receding into blue haze, with a huge illuminated NVIDIA emblem on the end wallGuides

Cloud GPU Pricing: When Renting Beats Buying a Card

H100s span $1.49 to $6.98 an hour across providers. Here's the break-even maths against buying, and the billing details that decide your real bill.

14 Aug5 min read
Cinematic local AI hardware illustration for: Why Are GPUs Used for AI? The Plain-English ExplainerGuides

Why Are GPUs Used for AI? The Plain-English Explainer

Why GPUs, not CPUs, run AI: parallel cores, matrix math and tensor cores, benchmarked with real VRAM and bandwidth specs from the RTX 4080 we render on daily.

12 Aug7 min read
Cinematic local AI hardware illustration for: Run Stable Diffusion Locally: Setup That Works in 2026Guides

Run Stable Diffusion Locally: Setup That Works in 2026

SD1.5 vs SDXL VRAM math, the ComfyUI desktop install, your first workflow, where checkpoints come from, and the errors that eat everyone's first afternoon.

12 Aug9 min read
Cinematic local AI hardware illustration for: Local LLMs for Coding: What Actually Works (2026)Guides

Local LLMs for Coding: What Actually Works (2026)

Which local coding models are real vs toys, wiring Ollama's OpenAI API into your editor, real latency numbers, and the privacy case for client code.

12 Aug9 min read
Cinematic AI audio production illustration for: Stable Audio 3 Local Setup: Run the Open Weights on Your GPUGuides

Stable Audio 3 Local Setup: Run the Open Weights on Your GPU

Stable Audio 3's open-weight models run on consumer hardware. The Hugging Face repos, the stable-audio-3 library, license limits, and what a 4080 can expect.

9 Aug6 min read
Cinematic local AI hardware illustration for: Poolside Laguna XS 2.1: A 33B Coding Model for One GPUGuides

Poolside Laguna XS 2.1: A 33B Coding Model for One GPU

Poolside's Laguna XS 2.1 activates just 3B of its 33B parameters per token, enough for real agentic coding on a single consumer GPU. Specs and setup.

7 Aug5 min read
Cinematic local AI hardware illustration for: Nemotron 3 Nano 4B: NVIDIA's Edge Model on Your RTX CardGuides

Nemotron 3 Nano 4B: NVIDIA's Edge Model on Your RTX Card

A 4B hybrid Mamba-Transformer model NVIDIA built for Jetson and RTX hardware, not data centers. Architecture, benchmarks, license, and the run command.

7 Aug5 min read
Cinematic AI video production illustration for: Wan 2.2 Animate: Character Animation on Your GPU (2026)Guides

Wan 2.2 Animate: Character Animation on Your GPU (2026)

Wan 2.2 Animate transfers a real performance onto any character image, motion and expressions included. ComfyUI setup, VRAM tiers, and mode choice.

3 Aug4 min read
Cinematic AI video production illustration for: Run MiniMax H3 Locally: Open Weights in ComfyUI (2026)Guides

Run MiniMax H3 Locally: Open Weights in ComfyUI (2026)

MiniMax H3's open weights are live. What you actually download, the 42.5GB ComfyUI footprint, the 12GB claim, and which parts stay hosted-only.

3 Aug5 min read
Cinematic AI video production illustration for: Run LTX-2 Locally: AI Video With Audio on 16GB (2026)Guides

Run LTX-2 Locally: AI Video With Audio on 16GB (2026)

LTX-2 generates video and synced audio in one pass on your own GPU. The real VRAM tiers, the 2.3 update, ComfyUI setup, and where 16GB cards hit limits.

3 Aug4 min read
Cinematic local AI hardware illustration for: NVIDIA RTX Spark for Local AI (2026): What It Actually RunsGuides

NVIDIA RTX Spark for Local AI (2026): What It Actually Runs

128GB unified memory, a 20-core Grace CPU, a Blackwell GPU on one chip. What RTX Spark can actually run locally, what it costs, and who should wait to buy.

2 Aug5 min read
Cinematic local AI hardware illustration for: Open-Weight AI Video Models You Can Actually RunGuides

Open-Weight AI Video Models You Can Actually Run

MiniMax H3 going open-weight restarts this argument. What runs on 8GB, what needs 24GB, and what the quantization actually costs you in quality.

1 Aug5 min read
Cinematic local AI hardware illustration for: Hunyuan 3 Local Guide: Can You Actually Run 295B at Home?Guides

Hunyuan 3 Local Guide: Can You Actually Run 295B at Home?

Tencent open-sourced Hunyuan 3 — 295B total, 21B active, Apache-2.0. We break down the real VRAM math, the GGUF quant path, and who should skip it entirely.

31 Jul6 min read
Cinematic local AI hardware illustration for: MiniMax M3 Locally: The 428B Model's Real Hardware CostGuides

MiniMax M3 Locally: The 428B Model's Real Hardware Cost

MiniMax's open-weight M3 posts frontier coding scores with a 1M-token context. What self-hosting a 428B MoE model takes, and who should just use the API.

27 Jul6 min read
Cinematic local AI hardware illustration for: DeepSeek V4 Local Guide: Real VRAM Needs in 2026Guides

DeepSeek V4 Local Guide: Real VRAM Needs in 2026

DeepSeek V4 shipped MIT-licensed open weights. But no stable Ollama build loads it yet — here's what actually runs locally, on what hardware, and what doesn't.

25 Jul5 min read
Cinematic local AI hardware illustration for: Kimi K3 Locally: The 2.8T Hardware Reality Check (2026)Guides

Kimi K3 Locally: The 2.8T Hardware Reality Check (2026)

Moonshot's Kimi K3 tops open-model charts, with weights landing July 27. What it actually takes to run 2.8T parameters at home — and what to run instead.

20 Jul4 min read
Cinematic local AI hardware illustration for: GLM-5.2 Local Setup: VRAM, Quants and Real Hardware PathsGuides

GLM-5.2 Local Setup: VRAM, Quants and Real Hardware Paths

Z.ai's GLM-5.2 leads every open-weights coding benchmark. Here's the honest VRAM math per quant, the three hardware paths that work, and who should bother.

20 Jul4 min read
Cinematic local AI hardware illustration for: Qwen3.6-27B Local Setup: A 27B Model That Beats a 397B OneGuides

Qwen3.6-27B Local Setup: A 27B Model That Beats a 397B One

Alibaba's Qwen3.6-27B fits on one RTX 4090 and edges its own 397B-parameter predecessor on coding benchmarks. Our Ollama setup, VRAM notes and first tokens/sec.

18 Jul5 min read
Cinematic AI video production illustration for: Run Wan 2.2 Locally: Free AI Video on Your Own GPU (2026)Guides

Run Wan 2.2 Locally: Free AI Video on Your Own GPU (2026)

Our exact local Wan 2.2 setup on an RTX 4080: GGUF quantization, the Triton + SageAttention + TeaCache speed stack, and the mistakes that waste a weekend.

17 Jul6 min read
Cinematic local AI hardware illustration for: Gemma 4 12B: Run Google's New Multimodal Model LocallyGuides

Gemma 4 12B: Run Google's New Multimodal Model Locally

Google DeepMind's Gemma 4 12B runs text, image, audio and video natively on a 16GB machine. What's confirmed at launch, and the Ollama setup.

17 Jul6 min read
Cinematic local AI hardware illustration for: Used RTX 3090 for AI in 2026: The 24GB Bargain GuideGuides

Used RTX 3090 for AI in 2026: The 24GB Bargain Guide

Why a used RTX 3090 is still the smartest VRAM-per-dollar buy for local AI, what 24GB actually unlocks, ex-mining risks, and the day-one tests that protect you.

16 Jul3 min read
Cinematic local AI hardware illustration for: Best GPU for Local AI in 2026: A VRAM-First GuideGuides

Best GPU for Local AI in 2026: A VRAM-First Guide

The GPU guide written from a rig that renders AI daily: why VRAM beats speed, what each budget tier really runs, and the used cards that embarrass new ones.

Updated 3 Aug4 min read
Cinematic local AI hardware illustration for: RTX 5070 Ti vs RTX 3090 for Local AI: 16GB New or 24GB Used?Comparisons

RTX 5070 Ti vs RTX 3090 for Local AI: 16GB New or 24GB Used?

The 5070 Ti's 16GB warranty versus a used 3090's 24GB ceiling. Bandwidth, tokens/sec, power and price compared — plus the one question that settles it.

Updated 1 Aug5 min read
Cinematic local AI hardware illustration for: RTX 3090 vs Radeon AI PRO R9700 (2026): 24GB or 32GB?Comparisons

RTX 3090 vs Radeon AI PRO R9700 (2026): 24GB or 32GB?

AMD's new $1,299 workstation card just launched with 32GB. We compare it against the used RTX 3090 we actually recommend most for local AI VRAM per dollar.

Updated 1 Aug5 min read
Cinematic local AI hardware illustration for: RTX 5090 vs 4090 for Local AI: Is 32GB Worth It?Comparisons

RTX 5090 vs 4090 for Local AI: Is 32GB Worth It?

The 5090's 32GB and 1,792 GB/s against the 4090's 24GB — what the bandwidth gap actually does to tokens per second, and when the used 4090 still wins.

25 Jul4 min read
Cinematic local AI hardware illustration for: Used RTX 3090 vs New RTX 4060 Ti 16GB for Local AI (2026)Comparisons

Used RTX 3090 vs New RTX 4060 Ti 16GB for Local AI (2026)

24GB used vs 16GB new, verified: bandwidth, tokens/sec on 8B-32B models, power draw, current pricing, and which model tier each card actually wins.

Updated 1 Aug6 min read
Two NVIDIA graphics cards side by side on a workshop bench under a large glowing NVIDIA sign, comparing 24GB and 16GB VRAM for local AIComparisons

Used RTX 3090 vs RTX 5060 Ti 16GB for Local AI (2026)

24GB used vs 16GB new, tested for local LLMs: memory bandwidth, real tokens/sec, power draw, and which model sizes each card can actually hold.

19 Jul4 min read
Cinematic local AI hardware illustration for: Local AI vs Cloud APIs: The Real Cost Math (2026)Comparisons

Local AI vs Cloud APIs: The Real Cost Math (2026)

We run both daily. When local hardware beats per-token APIs, when cloud wins, and the break-even math nobody shows: electricity, amortization, rental.

17 Jul3 min read
Cinematic local AI hardware illustration for: Ollama vs LM Studio (2026): Same-Hardware VerdictComparisons

Ollama vs LM Studio (2026): Same-Hardware Verdict

We run both on the same 16GB RTX 4080. Where Ollama's one-command workflow wins, where LM Studio's GUI and offload sliders win, and where we keep both.

Updated 27 Jul4 min read

More in Local AI