Together AI
thecontinuitylayer.com/tools/gpu-hosting/together-aiTeams that want a dedicated, managed endpoint for an open model without running servers
GPU cloud · Together AI
Together AI runs open models through a per-token API, and also rents GPUs in two ways: dedicated endpoints, where Together hosts a model you choose on GPUs reserved for you, and GPU clusters you control yourself.
Dedicated endpoints remove most of the setup work of serving an open model, at a higher hourly price than a bare GPU. Its published per-hour rates differ between the pricing page, docs and announcements, and reserved rates are lower for longer terms, so confirm the current figure before committing.
Specs
| Price per hour | GPU clusters: H100 $3.99/hr, B200 $8.19/hr per GPU on demand |
|---|---|
| GPUs | NVIDIA H100, H200, B200 and GB200 as dedicated endpoints or self-service clusters |
| Cluster H100 | $3.99 per GPU-hour on demand |
| Cluster B200 | $8.19 per GPU-hour on demand |
| Reserved H100 | From $3.19 per GPU-hour, by term |
| Dedicated endpoints | Billed per GPU-hour per replica |
| Other products | Per-token serverless API, fine-tuning |
Good for
- Managed hosting of an open model
- Moving from a per-token API to dedicated capacity
- Short cluster reservations
Not for
- The cheapest possible GPU hour
- Consumer-card workloads
See also
- Novita AI · Developers who want both an open-model API and rentable GPUs from one account
- GPU Droplets · Developers already on DigitalOcean who want a GPU server next to their app
- Hyperbolic · Developers who want on-demand data-center GPUs without a contract
We last reviewed Together AI on 31 October 2026. Prices and specs change; check together.ai before you buy or sign up. Figures we could not confirm on the vendor's own page read “not published” or “varies”.
Vendor source: https://www.together.ai/pricing