Luminal vs General Compute

Luminal compiles models into faster GPU code. General Compute gets its speed by moving decode onto ASICs from Cerebras and others.

By The Subconscious Team · Updated October 27, 2026

Luminal vs General Compute: key differences

Luminal's open-source compiler lowers models to native kernels ahead of time, and it reports GPT-OSS 120B at 36K tokens per second of aggregate throughput on 8 H100s against 26K for vLLM. Luminal Cloud is in early access, with DeepSeek V4.1 Flash at $0.1425 in and $0.57 out per million tokens and any PyTorch or Hugging Face model compiled to an endpoint. General Compute runs prefill on GPUs and decode on ASICs, and in its own benchmark measured over 1,000 output tokens per second on GPT-OSS 120B, which it sells at $0.21 in and $0.79 out. Both figures are self-reported and measure different things.

Pick Luminal for high-throughput serving of a custom model on your own GPUs, or a compiler you can inspect before buying. Pick General Compute for fast output on hosted open models, or reserved ASIC racks with a tokens-per-second SLA.

What Luminal and General Compute do

Luminal

Luminal builds an inference compiler. Where vLLM and SGLang interpret a model at runtime, Luminal compiles it ahead of time into native kernels for GPUs and ASICs. Models get lowered to a small graph of 15 primitive ops, and the compiler searches over fusion, tiling, memory and scheduling choices instead of relying on hand-written rules, which it says can find optimizations like Flash Attention on its own. The compiler is open source in Rust under Apache 2.0 or MIT, runs on CUDA and Metal with ROCm on the roadmap, and works as a torch.compile backend.

Example models: DeepSeek V4.1 Flash, Ideogram 4.5

Full Luminal profile

General Compute

General Compute is a San Francisco inference cloud built on alternative chips. It splits each request across hardware, with NVIDIA GPUs handling prefill and ASICs from Cerebras, SambaNova, Positron AI and d-Matrix handling decode. In its own open benchmark on GPT-OSS 120B it measured 1,025 to 1,234 output tokens per second against 100 to 121 on an unnamed GPU cloud, and up to 16.2x faster end to end. Founders Finn Puklowski and Jason Goodison raised a $15M seed led by FUSE, an up to $400M debt facility from Upper90 and, on October 27, 2026, a $100M Series A co-led by Blackbird and Abstract.

Example models: MiniMax M2.7, DeepSeek V3.2, GPT-OSS 120B

Full General Compute profile

Should you choose Luminal or General Compute?

Luminal

Choose Luminal for

Compiling custom models for owned GPUs
Open-source inference compiler
On-prem license with SLAs

General Compute

Choose General Compute for

Fast output on hosted open models
Plans from pay as you go to $1,000 a month
Reserved ASIC capacity

Luminal vs General Compute at a glance

Attribute	Luminal	General Compute
Model access	Small hosted catalog plus bring your own weights	Open weights
Flagship models	DeepSeek V4.1 Flash, Ideogram 4.5	MiniMax M2.7, DeepSeek V3.2, GPT-OSS 120B
Speed	36K tok/s on GPT-OSS 120B, 8xH100 (vendor)	1,000+ tok/s on GPT-OSS 120B (vendor)
Price	$0.1425 in, $0.57 out (DeepSeek V4.1 Flash)	$0.21–$0.28 in, $0.38–$1.20 out per 1M
Customization	Compiles any PyTorch or HF model	Custom LoRA and fine-tuned checkpoints
Deployment	Serverless (early access), on-prem license	Serverless API, dedicated racks, bare metal
Long context	Unknown	192K (MiniMax M2.7)

Frequently asked questions

What is the difference between Luminal and General Compute?

Luminal compiles models into faster GPU code. General Compute gets its speed by moving decode onto ASICs from Cerebras and others.

When should I choose Luminal over General Compute?

Compiling custom models for owned GPUs; Open-source inference compiler; On-prem license with SLAs.

When should I choose General Compute over Luminal?

Fast output on hosted open models; Plans from pay as you go to $1,000 a month; Reserved ASIC capacity.

Is Luminal or General Compute cheaper?

Luminal: $0.1425 in, $0.57 out (DeepSeek V4.1 Flash). General Compute: $0.21–$0.28 in, $0.38–$1.20 out per 1M. The cheaper choice depends on the model and workload.

Related comparisons

Running long-horizon agents?

If your agents run past 200K tokens, compare both against Subconscious. Our inference stack treats a long-horizon trace as the primary workload, so speed, cost, and accuracy hold up deep into the trace.

Subconscious vs Luminal Subconscious vs General Compute