Cerebras
Inference
The world’s fastest inference -
70x faster than GPU clouds,
128K context, 16-bit precision.
Latest Announcements
CePO: Empowering Llama with Reasoning using Test-Time Compute
Cerebras is proud to introduce CePO (Cerebras Planning and Optimization), a framework that adds sophisticated reasoning capabilities to the Llama family of models. Through test-time computation techniques, we enable Llama to tackle complex reasoning tasks with unprecedented accuracy. While models like OpenAI o1 and Alibaba QwQ have shown how additional computation at inference time can dramatically improve problem-solving capabilities [1], we’re now bringing these advances to Llama – the world’s most popular open-source LLM.
Cerebras Demonstrates Trillion Parameter Model Training on a Single CS-3 System
SUNNYVALE, CA AND VANCOUVER — December 10, 2024 – Today at NeurIPS 2024, Cerebras Systems, the pioneer in accelerating generative AI, today announced a groundbreaking achievement in collaboration with Sandia National Laboratories: successfully demonstrating training of a 1 trillion parameter AI model on a single CS-3 system. Trillion parameter models represent the state of the art in today’s LLMs, requiring thousands of GPUs and dozens of hardware experts to perform. By leveraging Cerebras’ Wafer Scale Cluster technology, researchers at Sandia were able to initiate training on a single AI accelerator – a one-of-a-kind achievement for frontier model development.
Cerebras Delivers Record-Breaking Performance with Meta's Llama 3.1-405B Model
Llama 3.1 405B now runs at 969 tokens/s on Cerebras Inference: Frontier AI now runs at instant speed. Last week we ran a customer workload on Llama 3.1 405B at 969 tokens/s – a new record for Meta’s frontier model. Llama 3.1 405B on Cerebras is by far the fastest frontier model in the world – 12x faster than GPT-4o and 18x faster than Claude 3.5 Sonnet. In addition, we achieved the highest performance at 128K context length and shortest time-to-first-token latency, as measured by Artificial Analysis.
Cerebras Sets a New Record in
Molecular
Dynamics
Cerebras CS-2 achieves 1 million simulation steps/s –
700x faster than the world’s #1 supercomputer
Award Winning Technology
Cerebras continues to be recognized for pushing the boundaries of AI
TIME
TIME
FORBES
FORTUNE
ai model services
You bring the data, we'll train the model
Whether you want to build a multi-lingual chatbot or predict DNA sequences, our team of AI scientists and engineers will work with you and your data to build state-of-the-art models leveraging the latest AI techniques.
high performance computing
The fastest HPC
accelerator on earth
With 900,000 cores and 44 GB of on-chip memory, the CS-3 completely redefines the performance envelope of HPC systems. From Monte Carlo Particle Transport to Seismic Processing, the CS-3 routinely outperforms entire supercomputing installations.
Models on Cerebras
The Cerebras platform has trained a huge assortment of models from multi-lingual LLMs to healthcare chatbots. We help customers train their own foundation models or fine-tune open source models like Llama 2. Best of all, the majority of our work is open source.
Latest Blog Posts
“Mayo Clinic selected Cerebras as its first generative AI collaborator for its large-scale, domain-specific AI expertise to accelerate breakthrough insights for the benefit of patients.”
Matthew Callstrom, MD, PhD
Medical Director for Strategy, Chair - Department of Radiology
"The Cerebras CS-2 is a critical component that allows GSK to train language models using biological datasets at a scale and size previously unattainable. These foundational models form the basis of many of our AI systems and play a vital role in the discovery of transformational medicines."
Kim Branson
SVP Global Head of AI and ML, GlaxoSmithKline
"Training which historically took over 2 weeks to run on a large cluster of GPUs was accomplished in just over 2 days — 52hrs to be exact — on a single CS-1. This could allow us to iterate more frequently and get much more accurate answers, orders of magnitude faster."
Nick Brown
Head of AI & Data Science, AstraZeneca
"Working with the Cerebras ML team we were able to train a new state-of-the-art large language model that outperforms models twice its size in a matter of weeks. Their AI expertise is second to none."
OpenTensor
The OpenTensor Foundation
"Cerebras allowed us to reduce the experiment turnaround time on our cancer prediction models by 300x, ultimately enabling us to explore questions that previously would have taken years, in mere months."
