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.

RRunyard Index

Runyard / GPU Index · LLM API Index →

GPU Rental Price Index

Last collected 2026-10-11 · 84 snapshots since 2025-06-04

H100 SXM by provider type

Median by provider type, per snapshot

HyperscalersNeocloudsLambda (archive, pre-live)Marketplaces

Median price by provider type, per snapshot. Hover for values. Lambda's archive history reaches back to June 2025 where it sold the GPU; everything else starts 2026-10-11.

Rent H100 SXM

You rent

GPU-hr

You pay · market median

$3.50

Cheapest listing: $1.68 at Latitude.sh

Rate1 GPU-hr ≈ $3.50

Per month, 720 h$2,520

Every H100 SXM provider →

Median of provider medians, on-demand, as collected 2026-10-11. Providers bill in their own currency; this is an estimate for comparison, not a quote.

H100 SXM median

$3.50/hr

B200 median

$6.99/hr

Providers tracked

GPUs priced today

32of 32

Consumer cards, mostly on marketplaces. Cheapest memory per dollar, with the caveats of shared, third-party hosts.

NVIDIA RTX 409024 GB GDDR6X · Ada Lovelace · ConsumerThe most-rented consumer card. 24 GB.Median$0.48/ GPU / hr

+9 providers 2

NVIDIA RTX 509032 GB GDDR7 · Blackwell · ConsumerThe consumer flagship. 32 GB and H100-class bandwidth at a fraction of the price.Median$0.81/ GPU / hr

+6 providers 3

NVIDIA RTX 309024 GB GDDR6X · Ampere · ConsumerTwo generations old, 24 GB, the cheapest way to rent that much memory.Median$0.22/ GPU / hr

+3 providers 4

NVIDIA RTX 408016 GB GDDR6X · Ada Lovelace · ConsumerAda consumer, 16 GB.Median$0.27/ GPU / hr

+2 providers 5

NVIDIA RTX 508016 GB GDDR7 · Blackwell · ConsumerBlackwell consumer, 16 GB.Median$0.38/ GPU / hr

+1 providers

Median = median of each provider's own on-demand median, so a provider with thirty instance sizes counts once. Logos are the cheapest providers for that GPU. Ranked by provider coverage, then price — never by our traffic.

More from the index

ProvidersEvery provider against the market, with history and price changes CompareH100 vs A100, H200 vs B200, RunPod vs Lambda…What compute costsOne table: median, cheapest, spot, per month, per million tokens By countryUSA, UK, India, Germany, France… each in its own currency

Providers

Provider	Type	How prices are collected
Amazon Web Services	hyperscaler	Public on-demand price list for Linux, us-east-1. Instance price divided by GPU count.
Microsoft Azure	hyperscaler	Azure Retail Prices API, Linux, eastus. Pay-as-you-go and Spot recorded separately. VM price divided by GPU count.
RunPod Secure Cloud	neocloud	Public GPU-type listing. Secure Cloud is RunPod's own datacentre capacity, quoted per GPU-hour.
RunPod Community Cloud	marketplace	Third-party hosts renting through RunPod. Quoted per GPU-hour; a listing may have no hosts behind it.
Lambda	neocloud	Public on-demand pricing page. Instance price divided by GPU count.
Vast.ai	marketplace	Public offer search, verified and rentable hosts only. Each offer is one observation; on-demand rate, not interruptible bids.
Oracle Cloud	hyperscaler	OCI's public price-list API (the cost estimator's feed), pay-as-you-go, per GPU-hour as listed. One price worldwide.
CoreWeave	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Massed Compute	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Civo	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Koyeb	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Oblivus	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
DataCrunch (Verda)	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
GPU.ai	marketplace	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Replicate	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Paperspace (DigitalOcean)	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Beam	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Latitude.sh	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Spheron	marketplace	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Runcrate	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Lyceum	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
fal.ai	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Enverge	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
HyperAI	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Packet.ai	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Cyfuture AI	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Vultr	hyperscaler	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Crusoe	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Hyperstack	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Nebius	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Together AI	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Modal	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Jarvislabs	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Thunder Compute	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
GMI Cloud	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Baseten	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
DigitalOcean GPU Droplets	hyperscaler	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Lightning AI	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
SaladCloud	marketplace	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Scaleway	hyperscaler	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Cerebrium	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Taiga Cloud	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Neysa	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
E2E Networks	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
TensorDock	marketplace	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
io.net	marketplace	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Akash	marketplace	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Northflank	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Qubrid	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Seeweb	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
OVHcloud	hyperscaler	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.
Denvr Dataworks	neocloud	Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded.

Questions

How much does it cost to rent a GPU right now?

As of 2026-10-11, across 49 providers: the H100 SXM has a market median of $3.50 per GPU-hour (cheapest listing $1.68 at Latitude.sh), the B200 $6.99, and the cheapest GPU-hour of any kind is the NVIDIA RTX A4000 at $0.16. Multiply by 720 for a month. Every GPU in the list has its own page with every provider's price.

What is the "median" and why not the cheapest?

The median is the middle of the market: each provider is first reduced to the median of its own prices for that GPU, so a provider with thirty instance sizes counts once, then the median is taken across providers. The cheapest listing is shown too, but it is usually a single marketplace host and can be gone in an hour. The median tells you whether a quote is high or low; the cheapest tells you the floor.

Where do the prices come from?

Every provider's own public pricing — their API where one exists, their pricing page where not — read every six hours. 49 providers, 1106 priced configurations today. No aggregator's data is licensed or scraped; the methodology page lists every source and the full calculation.

Why is the same GPU $2 at one provider and $11 at another?

Three markets share one name. Hyperscalers price the instance — CPU, RAM, network and the ecosystem — so the GPU is a fraction of it. Neoclouds price the GPU. Marketplaces are independent hosts competing on price with consumer-grade guarantees. The chart at the top stacks the three so the premium is visible rather than hidden.

How often does it update, and is there history?

Every six hours, and every snapshot is kept — 84 so far, since 2025-06-04. Each GPU page charts its median over time by billing type, each provider page charts its own prices, and price changes between snapshots are logged. Nothing is interpolated: a gap in collection is a gap on the chart.

Can I see prices in my currency?

Yes — the currency picker in the top bar converts every figure at the rate captured with the latest snapshot, and each country page prices the whole market in that country's currency against the global median.

Want to know which of these can actually hold a specific model? Every GPU page shows what fits at 4, 8 and 16-bit — and what the set costs per hour.

Start with the H100 →