Together AI Launches On-Demand Dedicated Endpoints for Scalable Inference
Action Required
Companies can now scale their AI applications more cost-effectively and with greater control, reducing the risk of performance bottlenecks and unexpected costs.
AI Impact Summary
Together AI is introducing on-demand Dedicated Endpoints, offering up to 43% lower pricing and unmatched price-performance for scaling AI inference. This new offering addresses the challenges of balancing flexibility and affordability for companies like BlackBox AI and DuckDuckGo, providing single-tenant GPU instances with full control and customizability, including support for custom models and autoscaling capabilities. This represents a significant shift from serverless deployments, particularly for applications requiring guaranteed performance and consistent latency.
Affected Systems
- Date
- Date not specified
- Change type
- capability
- Severity
- high