Templates Models

Build

AI Studio Notebooks GPU VMs ComfyUI

Train

Interactive training Experiment logging Training clusters PyTorch Lightning

Inference

Model APIs Dedicated models Deploy custom models Managed inference

GPUs Solutions Docs Pricing

Model APIsAccess top AI models instantly.
40 million free tokens to start.Dedicated modelsRun models on dedicated instances with instant autoscaling.Deploy a custom modelTurn your container into a scalable API in one click. Run on any cloud.

Top AI modelsBring AI into your product today. Production-grade models that scale with you.

Nemotron 3 Ultra

MiniMax M2.5 Deepseek V4 Pro

GPT-5.5 Gemma 4

Claude Fable 5 View all

AI StudioBuild and run AI in a collaborative GPU workspace. Use any IDE via SSH.NotebooksExperiment and analyze data in persistent GPU notebooks with AI help.GPU VMsLaunch raw GPU machines with full SSH access and persistent storage.

Build with appsRun AI apps in the cloud with zero setup.

ComfyUI OpenClaw

Interactive trainingCollaborative GPU workspace where AI agents help train and optimize models.Training clustersManaged frontier-grade GPU clusters with SLURM, K8s or multi-cloud LEC.PyTorch LightningMake large-scale PyTorch training simple. Train any model of any size.

Customize your training stackFor teams that need deep control, expert optimizations, and high-performance training.

Multi-node Faster data loading Batch Jobs

Thunder LitLogger

Build and test with GPUsInteractive workspace with agents to build & debug before going live.Deploy your own modelTurn your container into a scalable API in one click. Run on any cloud.Expert-managed inferenceOptimized and operated by Lightning’s PyTorch experts.

Customize your inference stackFor teams that need deeper control, custom pipelines, or full infrastructure control.

Agents (LitAI)Lightning deploy Batch Jobs Pipelines

On-demand GPUsRun VMs with direct ssh across 8+ clouds.Managed GPU clustersReserved clusters with SLURM, K8s. Flexible terms.GPU marketplaceGet the best rates and GPU availability on 8+ clouds.

Live GPU prices

L4$0.48T4$0.19L40S$2.89A100$1.55H200$4.33B200$12.54L4$0.48T4$0.19L40S$2.89A100$1.55H200$4.33B200$12.54

EnterpriseGo from prototypes to production AI securely.Frontier labsFully managed dedicated GPU clusters. Expert support.Model optimizationsFaster training and inference by the PyTorch experts.

AcademiaAccelerate research in universities and labs.

Customer stories

All docsDocs hub for the full Lightning ecosystem.Lightning AI StudioCloud workspace where AI helps you build and ship.PyTorch LightningPretrain and finetune any model of any size.

Learn

Tutorials Blog Community Courses Events OSS program About LitLogger

Pricing overviewExplore plans, GPU rates, model API prices, and FAQs.EnterpriseCustom contracts, dedicated support, and volume pricing.Frontier labsCompetitive prices for GPU clusters and 24/7 support.

Startups

All startups YC startups NVIDIA inception NVIDIA connect

Academia

Students Classrooms Research labs Universities

Request demoLog in Start free Open app

LitServe

Platform

Lightning AIDevelop, train and deploy on Lightning AI Python SDKProgrammatic access to the Lightning AI platform CLIAccess the Lightning AI platform from the command line

Open Source Libraries

PyTorch LightningFinetune and pretrain AI models. Focus on science, not engineering.Lightning FabricScale foundation models to 1000s of GPUs with expert-level control TorchMetrics90+ PyTorch metrics, optimized for distributed training ThunderCompiler for PyTorch: Make PyTorch models faster!LitLoggerExperiment tracking built into Lightning AI LitDataStream and process datasets at scale LitServeEasily serve AI models — Lightning fast LitAILLM router and minimal agent framework in one LitGPT20+ high-performance LLMs with recipes to pretrain, finetune and deploy at scale.

Home Install Quick start Speed up serving by 200x Benchmarks Advantages Migrate from FastAPI to LitServe FAQ Examples

Async and concurrency Authentication Autoscaling Batching Callbacks, middleware Deploy on cloud Dockerization Full control GPU inference Health check Hosting methods Logging and monitoring Model Context Protocol (MCP) servers Multiple API endpoints on one port Multiple models one server OpenAI spec OpenAI embedding Request, response format Serve LLMs Streaming Variable precision

LitAPI LitServer

Logging and Debugging ML API testing with pytest Programmatic shutdown Types of parallelism Use external APIs

Docs/LitServe/How to

How to

Learn how to perform common tasks in LitServe with clear, step-by-step instructions.

Was this page helpful?

Thanks for the feedback.

What went wrong?

Inaccurate or out of dateHard to understandMissing informationHard to find things / poorly organizedCode sample didn't workSomething is broken (link, image, layout)

Share with AI

Open in ChatGPT Open in Claude

Was this helpful?

Thanks for the feedback.

What went wrong?

Inaccurate or out of dateHard to understandMissing informationHard to find things / poorly organizedCode sample didn't workSomething is broken (link, image, layout)

↑↓ Navigate↵ Open⌘↵ New tabesc Close