# General Compute

> An inference cloud that runs decode on Cerebras, SambaNova and other ASICs instead of GPUs.

Canonical: https://www.subconscious.dev/providers/general-compute · By The Subconscious Team · Updated October 27, 2026

- Founded: 2025
- Example models: MiniMax M2.7, DeepSeek V3.2, GPT-OSS 120B
- Website: https://www.generalcompute.com

## Overview

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.

Its serverless API is OpenAI-compatible and runs in us-west-2. The catalog is MiniMax M2.7 with 192K context at $0.28 in and $1.20 out per million tokens, DeepSeek V3.2 with 32K context at $0.25 in and $0.38 out, and DeepSeek V3.1 and GPT-OSS 120B with 128K context at $0.21 in and $0.79 out. Plans run from pay as you go to $50 and $1,000 a month tiers with higher rate limits, and teams can bring LoRA or fully fine-tuned checkpoints behind a private model ID. The bigger business is dedicated capacity: reserved racks, bare metal or a guaranteed tokens-per-second SLA, priced per contract, with more than 50 MW in the pipeline and the first major deployments going live in Q1 2027.

## Upsides

- Fast decode on open models without being tied to one chip vendor.
- Published, reproducible benchmark code.
- Self-serve API with low prices, plus dedicated racks and tokens-per-second SLAs for large buyers.

## Core use cases

- Agentic coding and other agent loops where decode speed sets the wait.
- Voice agents and real-time chat.
- Teams that want reserved ASIC capacity under one contract.

## Downsides

- Small catalog of older open models, with DeepSeek V3.2 capped at 32K context.
- One region and OpenAI format only, with no Anthropic Messages endpoint.
- Most dedicated capacity does not go live until Q1 2027.

## At a glance

| Attribute | Value |
|---|---|
| Model access | Open weights |
| Flagship models | MiniMax M2.7, DeepSeek V3.2, GPT-OSS 120B |
| Speed | 1,000+ tok/s on GPT-OSS 120B (vendor) |
| Price | $0.21–$0.28 in, $0.38–$1.20 out per 1M |
| Customization | Custom LoRA and fine-tuned checkpoints |
| Deployment | Serverless API, dedicated racks, bare metal |
| Long context | 192K (MiniMax M2.7) |

## FAQ

### What is 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.

### What is General Compute best for?

Agentic coding and other agent loops where decode speed sets the wait; Voice agents and real-time chat; Teams that want reserved ASIC capacity under one contract.

### How much does General Compute cost?

General Compute pricing at a glance: $0.21–$0.28 in, $0.38–$1.20 out per 1M. Rates change often, so check General Compute's pricing page before committing.

### How much context does General Compute support?

General Compute's long-context support: 192K (MiniMax M2.7).

### What are the downsides of General Compute?

Small catalog of older open models, with DeepSeek V3.2 capped at 32K context; One region and OpenAI format only, with no Anthropic Messages endpoint; Most dedicated capacity does not go live until Q1 2027.

### What are the best alternatives to General Compute?

Common alternatives include Subconscious, OpenAI, Anthropic, Google Vertex AI, Amazon Bedrock. Each has a head-to-head comparison with General Compute on this site.

## Comparisons

- [Subconscious vs General Compute](https://www.subconscious.dev/compare/subconscious-vs-general-compute.md)
- [OpenAI vs General Compute](https://www.subconscious.dev/compare/openai-vs-general-compute.md)
- [Anthropic vs General Compute](https://www.subconscious.dev/compare/anthropic-vs-general-compute.md)
- [Google Vertex AI vs General Compute](https://www.subconscious.dev/compare/google-vertex-vs-general-compute.md)
- [Amazon Bedrock vs General Compute](https://www.subconscious.dev/compare/aws-bedrock-vs-general-compute.md)
- [Together AI vs General Compute](https://www.subconscious.dev/compare/together-ai-vs-general-compute.md)
- [Fireworks AI vs General Compute](https://www.subconscious.dev/compare/fireworks-vs-general-compute.md)
- [Baseten vs General Compute](https://www.subconscious.dev/compare/baseten-vs-general-compute.md)
- [Groq vs General Compute](https://www.subconscious.dev/compare/groq-vs-general-compute.md)
- [Cerebras vs General Compute](https://www.subconscious.dev/compare/cerebras-vs-general-compute.md)
- [DeepInfra vs General Compute](https://www.subconscious.dev/compare/deepinfra-vs-general-compute.md)
- [Hugging Face Inference Providers vs General Compute](https://www.subconscious.dev/compare/hugging-face-vs-general-compute.md)
- [Modal vs General Compute](https://www.subconscious.dev/compare/modal-vs-general-compute.md)
- [Cloudflare Workers AI vs General Compute](https://www.subconscious.dev/compare/cloudflare-workers-ai-vs-general-compute.md)
- [xAI vs General Compute](https://www.subconscious.dev/compare/xai-vs-general-compute.md)
- [Mistral AI vs General Compute](https://www.subconscious.dev/compare/mistral-ai-vs-general-compute.md)
- [DeepSeek vs General Compute](https://www.subconscious.dev/compare/deepseek-vs-general-compute.md)
- [Moonshot AI vs General Compute](https://www.subconscious.dev/compare/moonshot-ai-vs-general-compute.md)
- [Z.ai vs General Compute](https://www.subconscious.dev/compare/z-ai-vs-general-compute.md)
- [Alibaba Cloud vs General Compute](https://www.subconscious.dev/compare/alibaba-cloud-vs-general-compute.md)
- [Meta vs General Compute](https://www.subconscious.dev/compare/meta-vs-general-compute.md)
- [Cohere vs General Compute](https://www.subconscious.dev/compare/cohere-vs-general-compute.md)
- [SambaNova vs General Compute](https://www.subconscious.dev/compare/sambanova-vs-general-compute.md)
- [Nebius vs General Compute](https://www.subconscious.dev/compare/nebius-vs-general-compute.md)
- [Crusoe vs General Compute](https://www.subconscious.dev/compare/crusoe-vs-general-compute.md)
- [fal vs General Compute](https://www.subconscious.dev/compare/fal-vs-general-compute.md)
- [Novita AI vs General Compute](https://www.subconscious.dev/compare/novita-ai-vs-general-compute.md)
- [Venice vs General Compute](https://www.subconscious.dev/compare/venice-vs-general-compute.md)
- [Parasail vs General Compute](https://www.subconscious.dev/compare/parasail-vs-general-compute.md)
- [Inference.net vs General Compute](https://www.subconscious.dev/compare/inference-net-vs-general-compute.md)
- [GMI Cloud vs General Compute](https://www.subconscious.dev/compare/gmi-cloud-vs-general-compute.md)
- [Thinking Machines vs General Compute](https://www.subconscious.dev/compare/thinking-machines-vs-general-compute.md)
- [Sail Research vs General Compute](https://www.subconscious.dev/compare/sail-research-vs-general-compute.md)
- [Morph vs General Compute](https://www.subconscious.dev/compare/morph-vs-general-compute.md)
- [Relace vs General Compute](https://www.subconscious.dev/compare/relace-vs-general-compute.md)
- [TypeSafe AI vs General Compute](https://www.subconscious.dev/compare/typesafe-ai-vs-general-compute.md)
- [StepFun vs General Compute](https://www.subconscious.dev/compare/stepfun-vs-general-compute.md)
- [Runware vs General Compute](https://www.subconscious.dev/compare/runware-vs-general-compute.md)
- [StreamLake vs General Compute](https://www.subconscious.dev/compare/streamlake-vs-general-compute.md)
- [Wafer vs General Compute](https://www.subconscious.dev/compare/wafer-vs-general-compute.md)
- [RunInfra vs General Compute](https://www.subconscious.dev/compare/runinfra-vs-general-compute.md)
- [Particle.AI vs General Compute](https://www.subconscious.dev/compare/particle-ai-vs-general-compute.md)
- [Luminal vs General Compute](https://www.subconscious.dev/compare/luminal-vs-general-compute.md)
- [Infron vs General Compute](https://www.subconscious.dev/compare/infron-vs-general-compute.md)
- [Inception vs General Compute](https://www.subconscious.dev/compare/inception-vs-general-compute.md)
- [Dolphin vs General Compute](https://www.subconscious.dev/compare/dolphin-vs-general-compute.md)
- [OpenScale vs General Compute](https://www.subconscious.dev/compare/openscale-vs-general-compute.md)

## Sources

- [General Compute](https://www.generalcompute.com/)
- [General Compute models and pricing](https://docs.generalcompute.com/models)
- [General Compute benchmarks](https://www.generalcompute.com/performance)
- [General Compute Series A](https://www.generalcompute.com/blog/general-compute-100m-series-a)
- [Has the hunt for AI compute uncovered the next Cerebras?, TechCrunch](https://techcrunch.com/2026/05/28/has-the-hunt-for-ai-compute-uncovered-the-next-cerebras/)

Pricing and model lineups change often; figures are a snapshot.