Cerebras Training Cloud
Managed training/fine-tuning cloud for models from 1B to tens-of-trillions of parameters without manual sharding or model-parallelism configuration, billed pay-per-hour or pay-per-model (Cerebras-run custom training).
Find alternatives to Cerebras Training Cloud
What it does
- Model training
Features 2
- Train and fine-tune at any scaleTrains and fine-tunes models from 1 billion to tens of trillions of parameters with the same code, with no sharding or model parallelism.
- Preconfigured training environmentsEnvironments come preconfigured with secure storage and built-in networking so training can start without infrastructure setup.
Sources
- Pricing
-
The vendor's own cloud page, which is this product's pricing page.
- Description
- What it does
- Constraints
-
Derived from the security facts on this page, from the same source.
- Security
-
Company-wide Cerebras trust center (SafeBase, script-rendered; read in the in-app browser). Only the hosted services (Inference, Training Cloud) are listed, not the WSE hardware.
- Sold within
-
Official documentation · 14 Sep 2026
Own page states two billing models directly, independent of the existing cerebras-inference record which covers inference only.
- Name
- URL
Also from Cerebras Systems
-
Cerebras Inference Model API
Hosted inference service that serves open-weight models on Cerebras wafer-scale hardware, billed per token.
-
Cerebras WSE-3 Turbo (WSE-3T) Hardware
Wafer-scale processor — now WSE-3 Turbo (WSE-3T), four trillion transistors and 900,000 AI cores — sold inside Cerebras' rack-scale systems: three WSE-3T processors power the CS-4, the successor to the WSE-3-based CS-3.
-
Cerebras CS-4 Hardware
Rack-scale AI system built on three WSE-3 Turbo wafers per system and the Nexus rack-scale platform, aimed at ultrafast inference for frontier models; Cerebras says first shipments begin this quarter.
-
Cerebras CS-3 Hardware
Wafer-scale AI system built around the WSE-3 chip for organisations that run their own AI and HPC supercomputer; scales up to 24-trillion-parameter models on a single logical device and clusters from 1 to 2048 systems.
Open Cerebras Systems in the directory
Something wrong here? Send a correction — quote this product id: cerebras-training.