Runyard is a free hardware-aware AI model browser. You enter your CPU, GPU, and VRAM and it instantly shows every local LLM that will run on your machine, ranked by speed and quality.

How much VRAM do I need to run local LLMs?

8GB of VRAM runs 7B models like Llama 3.1 8B and Mistral 7B at Q4 quantization. 16GB unlocks 13B models. 24GB lets you run Mixtral 8x7B and Llama 3 70B at lower quantization.

What is the best local LLM for my GPU?

Use Runyard at www.runyard.dev — enter your GPU and VRAM and the Model Radar will rank every compatible LLM for your exact hardware, showing estimated tokens per second for each model.

Can I run Llama 3 locally?

Yes. Llama 3.1 8B at Q4 runs on any 8GB VRAM GPU. Llama 3.1 70B needs around 40GB VRAM at Q4, or an Apple Silicon Mac with 64GB+ unified memory.

Quantization Picker | Runyard Tools

P-03

Model size

Quantization

GPU or unified memory

System RAM

Context target

Quality preference

Speed preference

TurboQuant KV compression

Live result

Q5_K_M recommended

P-03

For 12 GB VRAM and a balanced preference, Q5_K_M is the cleanest first choice.

Fit-first

Q4_K_MQ5_K_M

Current preference

Quality-first

Q6_KQ5_K_M

Possible later

How It Works

3 inputs. Instant results.

01

Set the scenario

Choose realistic hardware, model, and context assumptions.

02

Read the result

The hero shows a working result instead of a decorative promo block.

03

Act on the outcome

Use the result to adjust fit, speed, quantization, or context.

Features

Everything that powers quantization picker.

01

Planning-first

Built to make local-AI decisions easier to reason about.

02

Local-AI focused

Built to make local-AI decisions easier to reason about.

03

Interactive hero

Built to make local-AI decisions easier to reason about.

04

Runyard design system

Built to make local-AI decisions easier to reason about.

05

GPU VRAM or unified memory

Grounded in the actual inputs and outputs this page is designed around.

06

Suggested quant range

Grounded in the actual inputs and outputs this page is designed around.

07

Friendly for first-time local users

Grounded in the actual inputs and outputs this page is designed around.

08

Standalone tool

Grounded in the actual inputs and outputs this page is designed around.

Spotlight

The differentiator behind quantization picker.

7B Q2_K on 8 GB

Fits~2.2 GBQuality loss trade

7B Q4_K_M on 8 GB

Unsure~4.7 GBRecommended start

7B Q8_0 on 8 GB

Might fit~7.1 GBVery tight

Visual comparison

Clarity

Fit

Actionability

Reading Results

How to read the output tiers.

Comfortable

<70%

Enough breathing room for normal use.

Tight

70%-95%

Should work, but overhead matters.

Borderline

95%-110%

Likely needs one tradeoff.

Too heavy

>110%

Time to step down.

Quick Reference

Common setups at useful defaults.

Scenario	Baseline	Result	Notes
Starter setup	7B / Q4 / 8K	Light local target	Good first benchmark
Balanced setup	8B / Q4 / 16K	Everyday sweet spot	Works for many users
Heavier setup	14B / Q5 / 16K	Quality-focused target	Needs stronger hardware
Stretch setup	32B / Q4 / 16K	Ambitious local target	Useful upper bound

* These are approximations for planning, not a promise of exact runtime behavior.

Benefits

Why people use quantization picker.

01

Faster decisions

It helps eliminate dead-end local AI choices before you download, benchmark, or configure too much.

02

Clearer tradeoffs

The page turns a raw estimate into something you can actually act on.

03

Useful on its own

The hero provides a working tool surface while the rest of the page explains what the output means.

FAQ

Questions people ask before using quantization picker.

What does quantization actually do to a model?

It reduces the bits used to store each weight, shrinking file size and VRAM at the cost of some precision. Q4_K_M uses ~4.5 bits per weight; F16 uses 16. Less bits = smaller, faster, slightly worse quality.

What is the difference between Q4_K_M and Q4_K_S?

K_M uses a medium-size importance matrix during quantization, preserving more detail in sensitive layers. K_S uses a smaller matrix. Q4_K_M is consistently recommended over K_S for better quality at the same size.

At what VRAM budget should I step from Q4 to Q5?

If you have 2–3 GB headroom after loading at Q4_K_M, Q5_K_M is worth trying. It improves reasoning and coding outputs noticeably at a moderate size increase — roughly 20% more VRAM.

When should I use Q2 or Q3?

Only when VRAM is severely constrained and you accept visible quality loss. Q2_K cuts model size nearly in half vs Q4 but degrades coherence on longer outputs. Use it as a last resort, not a first choice.

Does quantization affect coding quality more than chat?

Yes. Reasoning and code generation are more sensitive to weight precision. For general chat, Q4_K_M is often indistinguishable from Q6. For complex multi-step code or math, Q5 or Q6 makes a real difference.

Is there a level that gives near-original model quality?

Q6_K gives near-lossless quality at ~6.5 bits per weight. Q8_0 is effectively lossless. Both cost significantly more VRAM than Q4 — mainly suitable for 24 GB+ GPUs where the headroom exists.

RUNYARD.DEV / Tools / Quantization Picker

Estimates on this page are directional and should be validated against your actual runtime and hardware.