Pioneering AI computing company Cerebras Systems has introduced a groundbreaking solution that will revolutionize AI inference. On August 27, 2024, the company announced the launch of Cerebras Inference, the world's fastest AI inference service. Cerebras Inference's performance metrics overshadow traditional GPU-based systems, offering 20 times the speed at a fraction of the cost, setting a new standard for AI computing.
Cerebras Inference is particularly adept at handling various AI models, especially rapidly evolving "large language models" (LLMs). Taking the latest Llama3.1 model as an example, its 8B version can process 1,800 tokens per second, while the 70B version handles 450 tokens. This speed is not only 20 times that of NVIDIA GPU solutions but also more competitively priced. Cerebras Inference starts at just 10 cents per million tokens, with the 70B version at 60 cents, offering a 100-fold increase in cost-effectiveness compared to existing GPU products.
Impressively, Cerebras Inference achieves this speed while maintaining industry-leading accuracy. Unlike other solutions prioritizing speed, Cerebras consistently operates within a 16-bit numerical domain, ensuring that performance improvements do not compromise the quality of AI model outputs. Mihai Seagul-Smith, CEO of Artificial Analysis Corp, noted that Cerebras achieved a record-breaking speed of over 1,800 output tokens per second on Meta's Llama3.1 model.
AI inference is the fastest-growing segment of AI computing, accounting for about 40% of the entire AI hardware market. The high-speed AI inference provided by Cerebras, akin to the advent of broadband internet, opens up new opportunities and heralds a new era for AI applications. Developers can leverage Cerebras Inference to build next-generation AI applications requiring complex real-time performance, such as intelligent agents and systems.
Cerebras Inference offers three reasonably priced service tiers: Free Tier, Developer Tier, and Enterprise Tier. The Free Tier provides API access with generous usage limits, ideal for a wide range of users. The Developer Tier offers flexible serverless deployment options, while the Enterprise Tier provides customized services and support for organizations with continuous loads.
At its core, Cerebras Inference utilizes the Cerebras CS-3 system, powered by the industry-leading Wafer Scale Engine 3 (WSE-3). This AI processor is unparalleled in scale and speed, offering 7,000 times the memory bandwidth of the NVIDIA H100.
Cerebras Systems not only leads in AI computing but also plays a crucial role in various industries including healthcare, energy, government, scientific computing, and financial services. By continuously advancing technological innovation, Cerebras is helping organizations across sectors tackle complex AI challenges.
Key Highlights:
🌟 Cerebras Systems offers 20 times the service speed at more competitive prices, ushering in a new era for AI inference.
💡 Supports a wide range of AI models, particularly excelling in large language models (LLMs).
🚀 Offers three service tiers, allowing developers and enterprise users to choose flexibly.