Pay per model or by the hour

Pay Per Hour

Share your AI workload with Cerebras and we’ll determine the time needed for experimentation and production model training. You’ll receive access to a secure, dedicated programming environment you can use to build your model.

Contact Sales

Pay Per Model

Let Cerebras’ ML experts do the work! Share your AI solution of interest and your dataset and our ML experts will thoroughly design and rapidly build a state-of-the-art generative AI solution for you.

See our AI Model Services

The easiest way to train or fine-tune large AI models

The Cerebras AI Model Studio is a purpose-built platform, optimized for training and fine-tuning large language models on dedicated clusters. It provides deterministic performance, requires no distributed computing headaches, and is push-button simple to start.

Models on Cerebras

Cerebras has trained a huge assortment of models from multi-lingual LLMs to healthcare chatbots. These models are available for training on our cloud.

llama 2

Foundation language model
7B-70B, 2T tokens
4K context


7B Foundation model that leverages
grouped-query attention,
coupled with sliding window attention


Bilingual Arabic + English model
13B, 30B Parameters
Available on Azure, G42 Cloud


Medical Q&A LLM
Fine-tuned from Llama2-70B
Scores 72% on USMLE


Massive multi-lingual LLM
176B parameters, 366B tokens
2k context


Foundation language model
40B, 1T tokens,
(Uses Flash Attention and Multiquery)


Foundation model trained
on 1T tokens of English
that uses ALiBi positioning method


Coding LLM
15.5B parameters, 1T tokens
8K context


Image generation model
33M-2B parameters
Adaptive layer norm


For NLP applications
Encoder-decoder model
60M-11B parameters


Trained for English + Code
7B Parameters, 1.3T Tokens
LLM360 Release


Foundational Language Model
100m - 13b parameters


BTLM-3B-8K fine-tuned for chat
3B parameters, 8K context
Direct Preference Optimization


Implements nanoGPT on Cerebras
Trains 175B+ models
565 lines of code

Key Benefits

Large models in less time

Train 1 billion to 24 trillion parameter models faster than the largest publicly available Cloud GPU instances

Simple & Easy to Use

Enjoy exa-scale performance with the simplicity of programming a single device


Competitive pricing options, including pay-by-model and pay-per-hour, that are tailored for your budget


Train your models from scratch or fine-tune open-source models with your data


Dependency free - Keep the trained weights for the models you build

Simple & Secure cloud operations

Simple onboarding: no DevOps required
Software environment, libraries, secure storage, networking configured and ready to go

Contact us to learn more