Freedom251 › Ai Development Tools
Together AI Dedicated Inference
Plans
| Plan | Billed and limits | Price |
|---|---|---|
| Dedicated Inference — NVIDIA HGX H100 (on-demand) | On-demand, pay as you go; billed per GPU per hour · Single-tenant GPU instances · guaranteed performance (no sharing) · custom models · autoscaling and traffic spike handling | $5.49 |
| Dedicated Inference — NVIDIA HGX B200 (on-demand) | On-demand, pay as you go; billed per GPU per hour · Single-tenant GPU instances · guaranteed performance (no sharing) · custom models · autoscaling and traffic spike handling | $8.99 |
| Dedicated Inference — reserved capacity | Contact sales · Reserved GPU capacity · pricing depends on hardware and configuration | On request |
As together.ai lists them, read 8 Oct 2026. The maker's own page has the final word.
Facts from Together AI Dedicated Inference's own pages
- API compatibility
- Dedicated and serverless inference use the same OpenAI-compatible API, so a serverless prototype can move to a dedicated endpoint by changing the endpoint. · together.ai, 8 Oct 2026
- Autoscaling and routing
- The product page lists SLO-driven autoscaling based on time to first token, latency, and throughput, along with least-loaded, session-affinity, and prefix-cache-aware routing. · together.ai, 8 Oct 2026
- Best suited for
- Together lists predictable or steady traffic, latency-sensitive applications, and high-throughput production workloads as best suited to Dedicated Model Inference. · together.ai, 8 Oct 2026
- Company mission
- Together AI describes its mission as designing a full-stack AI platform powered by systems research to help teams ship faster, scale reliably, and achieve superior unit economics. · together.ai, 8 Oct 2026
- Compliance and data residency
- Together says customers can choose a deployment region for data residency and identifies SOC 2 Type II and ISO 27001 as supporting certifications. · together.ai, 8 Oct 2026
- Custom models
- Customers can deploy custom or fine-tuned models from Hugging Face or their own S3 storage through the UI or CLI. · together.ai, 8 Oct 2026
- Deployment
- Customers select a target model and hardware configuration to deploy a production-ready endpoint, and Together says it can be live in minutes without deep infrastructure expertise. · together.ai, 8 Oct 2026
- Founded
- 2022 · together.ai, 28 Sept 2026
- Headquarters
- San Francisco, California, United States · together.ai, 28 Sept 2026
- Pricing basis
- Dedicated Inference is billed per GPU-hour based on GPU type and replica count; the pricing page lists on-demand rates for HGX H100 and HGX B200, while reserved rates require contacting sales. · together.ai, 8 Oct 2026
- Rollouts
- Customers can test updated deployments with shadow traffic or A/B testing, and Together says it manages the rollout without taking the serving endpoint down. · together.ai, 8 Oct 2026
- Security and privacy
- Dedicated deployments use single-tenant, isolated GPUs; Together says traffic and data are not shared and that it does not train on customer prompts or model weights. · together.ai, 8 Oct 2026
- Service
- Dedicated inference runs a model on GPUs reserved exclusively for a customer's workload, providing consistent latency, predictable cost, and control over the model and hardware. · together.ai, 8 Oct 2026
- Support
- The product FAQ directs users to contact Together's team for help, and the product page offers a Talk to Sales option. · together.ai, 8 Oct 2026
Where Together AI Dedicated Inference stands on Freedom251
- Ai Development Tools#23 of 26