Inference at the
speed of light.

Luminal compiles AI models to give you the fastest, highest throughput inference in the world.

GitHub
The Compiler

Compiled inference, not interpreted

Unlike runtime inference engines that interpret models dynamically, Luminal compiles your model ahead of time into optimized native code for GPUs and ASICs, eliminating every layer of overhead.

Compilation Pipeline
MODEL
PyTorch / HF
GRAPH IR
Dataflow graph
OPTIMIZE
Fuse & schedule
GPU / ASIC
Execute

Graph-Level IR

01

Models are lowered to a minimal graph intermediate representation, a pure dataflow graph with no framework overhead.

Hardware-Aware Optimization

02

The compiler applies fusion, tiling, memory planning, and scheduling passes tuned for each target, GPUs and ASICs.

Zero-Overhead Codegen

03

Final code is emitted directly to GPU kernels or ASIC instructions with no excess runtime overhead.

The Inference OS

Hyperscale Inference OS

Luminal dynamically schedules and load-balances inference workloads at any scale, from single accelerators up to large clusters of heterogeneous compute nodes, minimizing latency and maximizing throughput by optimizing inference topologies on-the-fly.

Compute Cluster
gpu
asic
cpu
GPU-0
GPU
FLUX-[dev]
94% UTIL
CPU-0
CPU
Preprocessing
45% UTIL
ASIC-0
ASIC
FLUX-[dev]
72% UTIL

Heterogeneous Compute

01

Inference across CPUs, GPUs, and ASICs deliver maximum throughput and superior TCO.

Dynamic Load Balancing

02

Continuously monitors utilization across every node and redistributes work in real time to eliminate bottlenecks and hotspots.

Lightning Quick Scaling

03

Nodes are dynamically booted and shutdown as workloads fluctuate, meeting peak loads without excess idle capacity.

Performance

Unmatched throughput

Our compiler-first approach eliminates runtime overhead entirely. Models compiled by Luminal consistently outperform existing inference engines by 2-3x on standard benchmarks.

3.2x
vs vLLM
<10ms
p99 latency
0%
overhead
Tokens/sec, GPT-OSS 120B, 8xH100 SXM
Luminal36k tok/s
TensorRT-LLM28k tok/s
vLLM26k tok/s
PyTorch3k tok/s
Deployment

Choose your deployment

Luminal Cloud

Managed serverless inference. Deploy in minutes, scale automatically, and only pay for what you use.

  • Serverless inference endpoints
  • Scale to zero capabilities
  • Automatic batching
  • Optimized compilation
  • Pay only for what you use
On-Prem Deployment

Run Luminal on your own infrastructure with dedicated support and enterprise-grade security.

  • Licensed cloud or on-prem deployment
  • Dedicated engineering support
  • Custom kernel optimization
  • Strict SLAs tailored to you
Get Started

Ready to accelerate your inference?

Get early access to the fastest AI inference platform in the world.