What can I request through Lilac Dedicated?

Lilac Dedicated supports requests for dedicated NVIDIA GPU virtual machines, bare-metal nodes, and multi-node clusters. Bare-metal and cluster requests may start from an eight-GPU node where appropriate, with the final configuration qualified by the Lilac team.

How does Lilac Serverless work?

Lilac Serverless provides OpenAI-compatible, per-token inference for supported models without requiring a GPU reservation.

How does Lilac Flex work?

Committed customers may opt in to release specific unused capacity windows. When Lilac matches a released window with approved demand, the customer can earn bill credits under the applicable agreement. Matching and credits are not guaranteed.

What is the Capacity Exchange?

The Capacity Exchange lets customers relist eligible reservations or take over an approved commitment from a verified customer. Any transfer is subject to demand, market pricing, customer verification, credit approval, provider approval, and applicable terms.

Backed by Y Combinator

The GPU cloud built for
fast-moving startups.

Reserve GPU VMs, bare-metal nodes, or private clusters.
Add capacity, release eligible windows, or transfer an approved commitment as plans change.

View GPU pricing Get started

Trusted by

Z.ai Y Combinator Osmosis Understudy Labs Saturn Cloud Eden AI NanoGPT Linum Infron Coral Bricks Floot

Why Lilac

Dedicated GPUs. Flexibility built in.

BuyCapacity from Lilac

Reserve GPU VMs, dedicated nodes, or multi-node clusters from Lilac. We provide the capacity, access, and support.

MonetizeUnused reservation windows

Use Lilac Flex to offer eligible unused windows to spot workloads. When demand is matched, the resulting credits reduce your GPU bill.

ResellEligible commitments

Relist an eligible reservation or take one over from a verified startup. Transfers depend on demand, pricing, and approval.

Products

One GPU Network.
Four ways to use Compute.

Dedicated clusters

Dedicated GPU infrastructure, directly from Lilac

Compare indicative pricing across GPU VMs, bare-metal nodes, and multi-node clusters. Lilac handles capacity, access, and support.

H100 to B300

Reserve H100s

Indicative pricing. Final rates depend on configuration and term.

*Estimate pricing

Serverless

Frontier models, simple per-token pricing

Use supported open models through an OpenAI-compatible API. No GPU reservation, no minimums, and no infrastructure to manage.

4+ modelsGet started

MiniMax M3

Speed

-tok/s

TTFT

Input /M

$0.28

Cache /M

$0.05

Output /M

$1.10

1MFP8

Kimi K2.6

Speed

-tok/s

TTFT

Input /M

$0.70

Cache /M

$0.20

Output /M

$3.50

262KINT4

GLM 5.2

Speed

-tok/s

TTFT

Input /M

$0.90

Cache /M

$0.27

Output /M

$3.00

524KFP8/NVFP4

Gemma 4 (31B)

Speed

-tok/s

TTFT

Input /M

$0.11

Cache /M

Output /M

$0.35

262KFP8

Subscriptions

Monthly inference credits that stretch further

Choose a flat monthly credit pool for supported models. Lower-cost capacity lets a predictable subscription cover more product usage.

Up to 12x valueGet started

Basic$10/moUp to $80 value Pro$30/moUp to $300 value Max$100/moUp to $1200 value

Batch

Run GPU jobs to completion

Submit a container image and a command. Lilac runs it on available GPU capacity with simple per-second pricing and no cluster to manage.

Private betaJoin private beta

Batch jobsPrivate beta

eval-suite-nightlyrunning

H100 $1.00/hr · 2.82 hrs · ~$2.82

train-baselinerunning

H200 $1.50/hr · 1.14 hrs · ~$1.71

data-processdone

H100 $1.00/hr · 0.46 hrs · ~$0.46

Commitment lifecycle

Commit without getting trapped.

Reserve what you need now. Use Lilac Flex to sell eligible downtime into spot demand, or transfer an approved commitment as plans change.

Reserve

Choose the capacity you need now.

Select the GPU, region, scale, and term. The final quote names the site, network, delivery date, and commercial terms.

Lilac Flex

Automatically sell eligible downtime.

Opt in to offer unused reservation windows to Lilac's spot workloads. When demand is matched, bill credits lower your effective GPU cost.

Transfer

Change the owner when plans change.

Relist an eligible commitment or take one over from a verified startup. Lilac qualifies the buyer and coordinates approval.

Talk to our team

Blog

Notes from the team

Updates, thinking, and technical deep-dives from the Lilac team.

View all posts

Lucas Ewing15 Jul 2026

Lilac partners with Saturn Cloud to expand GPU capacity for Token Factory

Saturn Cloud can scale Token Factory model serving and per-token inference on Lilac capacity without requiring customers to reserve new hardware.

Read details...

Lucas Ewing7 May 2026

We're partnering with MiniMax to bring M2.7 to Lilac

We are partnering with MiniMax to bring commercially licensed MiniMax M2.7 access to Lilac.

Read details...

Lucas Ewing7 May 2026

How to keep frontier open weights viable

Why Lilac supports open-weight licensing, and why commercial rights can help more frontier models stay open.

Read details...

Lucas Ewing1 May 2026

Kimi K2.6 is live on Lilac

Kimi K2.6 is now available on Lilac with OpenAI-compatible chat completions, 262K context, cache-read pricing, and no commitments.

Read details...

View all posts

GPU cloud for startups

The compute is already on.
Put it to work.

View GPU pricing Get started

The GPU cloud built forfast-moving startups.

Dedicated GPU infrastructure, directly from Lilac

Reserve H100s

Frontier models, simple per-token pricing

MiniMax M3

Kimi K2.6

GLM 5.2

Gemma 4 (31B)

Monthly inference credits that stretch further

Run GPU jobs to completion

Choose the capacity you need now.

Automatically sell eligible downtime.

Change the owner when plans change.

Lilac partners with Saturn Cloud to expand GPU capacity for Token Factory

We're partnering with MiniMax to bring M2.7 to Lilac

How to keep frontier open weights viable

Kimi K2.6 is live on Lilac

The GPU cloud built for
fast-moving startups.