Digital Event Horizon
In a significant development, Together AI has introduced dedicated GPU infrastructure, known as Together Reasoning Clusters, designed specifically for large-scale and low-latency inference of reasoning models. This move aims to support companies integrating reasoning models at scale, offering high-performance infrastructure, customized optimization, and security.
The Together AI has expanded its infrastructure to support the growing demand for its DeepSeek-R1 reasoning model. The company is introducing dedicated GPU infrastructure, known as Reasoning Clusters, designed for large-scale and low-latency inference of reasoning models. Together Reasoning Clusters offer instant scalability, flexible pricing, and higher rate limits, making them an attractive option for companies with variable workloads. The clusters are optimized for speed, cost efficiency, and security, with customized optimization for each client's traffic profile. The move positions Together AI as a leader in the field of reasoning models for enterprise applications.
In a recent update to its offerings, Together AI has announced the expansion of its infrastructure beyond its ultra-fast serverless API for DeepSeek-R1. This move is part of an effort to support companies that are integrating reasoning models into their production applications at scale.
According to the company, the demand for DeepSeek-R1 has been on the rise since its launch, with a surge in companies looking to leverage the power of reasoning models for various applications. To meet this growing demand, Together AI is introducing dedicated GPU infrastructure, known as Together Reasoning Clusters, which are specifically designed to handle large-scale and low-latency inference of reasoning models.
The introduction of Together Reasoning Clusters comes on the heels of the company's announcement of its fast serverless API for DeepSeek-R1. This API offers the fastest inference available for the full DeepSeek-R1 671B model, outperforming all other providers, including the model creator's own API. It also provides instant scalability, flexible pricing, and higher rate limits, making it an attractive option for companies with variable workloads.
The new offering is particularly appealing to organizations that require consistent, low-latency performance to meet production demands. Together Reasoning Clusters are designed to provide this level of performance, with dedicated infrastructure optimized for speed, cost efficiency, and security. These clusters run with the proprietary Together Inference stack, which has been shown to be 2.5x faster than open-source engines like SGLang.
Furthermore, Together AI's Reasoning Clusters offer customized optimization for each client's traffic profile. The company's expert research team fine-tunes key parameters such as batch sizes, prompt caching, and resource allocation to ensure that the workload achieves the optimal balance between throughput and latency based on its specific needs.
This move is a significant development in the field of reasoning models, which have been gaining traction in recent years due to their ability to enable machines to reason and learn like humans. Companies such as Together AI are playing a crucial role in making these technologies accessible to a wider range of organizations, from startups to enterprises.
With the introduction of its Reasoning Clusters, Together AI is positioning itself as a leader in the field of reasoning models for enterprise applications. The company's commitment to providing high-performance infrastructure, customized optimization, and security is likely to appeal to companies looking to integrate reasoning models into their production workflows.
Related Information:
https://www.together.ai/blog/deploy-deepseek-r1-at-scale-fast-secure-serverless-apis-and-large-scale-together-reasoning-clusters
Published: Mon Feb 17 20:52:25 2025 by llama3.2 3B Q4_K_M
