Cerebras Unveils CS-4
Cerebras Unveils CS-4
Cerebras has introduced the new CS-4 server, which allows models to run at twice the speed of previous versions.
Cerebras Unveils CS-4
The new Cerebras CS-4 server enables models to run at speeds twice as fast as previous versions, allowing for 10T models to be executed at speeds exceeding 1000 tokens per second. Each server is equipped with three WSE 3 Turbo chips, which are an overclocked version of the original WSE 3.
The company has doubled the frequency from 1.4 GHz to 2.8 GHz, which has increased memory bandwidth and FLOP/s, resulting in improved model speed. However, the memory on the chip remains at 44 gigabytes, so several dozen CS-4 servers are needed for the promised 10T models.
A unique feature of the CS-4 is its modularity: each chip is housed in a separate module that connects to a rack responsible for power and cooling. This simplifies installation, as racks can be set up in advance, and chips can be connected when they are ready. Additionally, this allows for future upgrades.
Why it matters
AnalysisThis server opens new possibilities for running large artificial intelligence models, which can significantly impact data processing speeds across various fields.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
GPT Image Generative Models Prompting Guide
OpenAI's gpt-image generative models are designed to produce production-quality visuals and fully controllable creative workflows. They are suitable for both professional design ta...
Claude Sonnet 4.6 vs GPT-5.5 vs Gemini 3.1 Pro: 2026 model comparison
Every few weeks, a new model comes out, and every time the same questions arise: should I switch? what changed what to pay for?
Thinking Machines Releases Inkling 2 Small: a 594B MoE Model That Can Fine-Tune Other Models
Thinking Machines released Inkling 2 Small, a 594B-parameter MoE model that leads among open models in several benchmarks and demonstrated autonomous fine-tuning of Qwen3.5-4B.
Community Discusses Anthropic Bans Over the Past Week
Comments mention users reporting accounts banned by Anthropic during the past week, with no details provided.