# Together GPU Clusters (slim)

> Together GPU Clusters rents multi-node NVIDIA H100, H200, B200 and B300 clusters with InfiniBand and shared storage, run as Kubernetes or Slurm, by the GPU-hour. Clusters are created and scaled from the console, the tg CLI, SDKs or REST API.

- Full: https://www.anchorterminal.com/tools/together-gpu-clusters.md (~7,800 tokens) · this version ~1,680 tokens · JSON https://www.anchorterminal.com/tools/together-gpu-clusters.json · canonical https://www.anchorterminal.com/tools/together-gpu-clusters
- Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · MCP: https://www.anchorterminal.com/mcp · Updated: 2026-11-07 · pricing https://www.anchorterminal.com/tools/together-gpu-clusters#pricing · reviews https://www.anchorterminal.com/tools/together-gpu-clusters#reviews
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)

**C · 58.1/100 · rank #764 of 1486 · #10 in GPU & serverless compute · not agent-ready · confidence medium**


Assessment: Multi-node H100 to B300 clusters on Kubernetes or Slurm, created and deleted headlessly through an OpenAPI-described API, a CLI with JSON output and two SDKs, at published GPU-hour prices. Clusters take 10 minutes to an hour to provision, can't be paused, and have no status component, SLA or idempotent create.

## Facts

- Kind: HTTP API · vendor: Together AI · category: GPU & serverless compute · legal entity: Together Computer, Inc. · provenance 83/100
- Endpoint: `https://api.together.ai/v1/compute/clusters` ()
- Auth: API key · pricing: Pay per use · x402: no · licence: Proprietary service under Together's terms of service. The Python SDK and CLI are Apache-2.0 and the OpenAPI document is MIT
- Benchmark probes: not run yet, so Performance is pending
- Live, from our pollers (not part of the score): up at 2026-11-07 18:21 UTC · uptime 30 days 100.0% (200 probes) · p50 491 ms · p95 857 ms over 24 hours
- Surface listed: Together GPU clusters (Instant Clusters) through the REST API at api.together.ai/v1/compute/clusters, the `tg clusters` CLI and the Python and TypeScript SDKs, read from docs.together.ai (overview, quickstart, concepts, management, automation, pricing, preemptible compute, API reference and changelog) and status.together.ai on 6 November 2026. No account was made and no cluster was created
- Scope: Multi-node GPU clusters for training and batch work. Serverless and dedicated inference, fine-tuning and Code Sandbox have separate Anchor Terminal listings
- Hardware: NVIDIA HGX H100 SXM (80 GB), H200 (141 GB), B200 (180 GB) and B300, in full nodes of 8 GPUs, with 3.2 Tb/s InfiniBand, local NVMe and a shared volume of at least 1 TiB
- Cluster types: Kubernetes, reached with kubectl, or Slurm on Kubernetes through Slinky, reached by SSH to a login node
- Capacity types: On-demand by the hour, reserved for 3 to 180 days, scheduled windows from a project capacity pool, and preemptible nodes (preview) with a five-minute drain
- Access: Project roles admin and editor. Only admins download the admin kubeconfig, which has cluster-admin rights. OIDC gives editors kubectl access under Kubernetes RBAC
- Health: Active and passive health checks, with repair recommendations that can be approved by hand or run fully automatically
- Lifecycle: Reserved clusters are decommissioned at the end of the reservation after emails 7 days, 3 days, 1 day and 4 hours before. Volumes outlive clusters and are deleted 72 hours after credit runs out
- Prices: H100 on-demand $3.99 per GPU-hour; H200 on-demand $5.99 per GPU-hour; B200 on-demand $8.19 per GPU-hour; B300 on-demand $9.99 per GPU-hour

- Scores: Reliability 38, Performance pending, Schema & documentation 92, Agent ergonomics 55, Security & auth 68, Payments & pricing 20, Task success pending, Maintenance & community 75, Transparency & trust 65 · total over the 7 assessed categories
- Why: Reliability, Read with the hosted lines. · Schema & documentation, Public OpenAPI 3.1 for /v1/compute/clusters and shared volumes, MIT, embedded in every reference page (25). · Agent ergonomics, Cluster objects are compact, and `--json` gives machine-readable output, but list calls take no field selection (15). · Security & auth, Project-scoped API keys, revocable with optional expiry, read as Together's other listings read the same key model (30). · Payments & pricing, No machine payment protocol (0). · Maintenance & community, Closed service. · Transparency & trust, Closed service under Together's terms, with Apache-2.0 SDKs, which don't raise the line for a closed API (15).
- Sources: 12, open questions: 4, both in the full twin
- Capabilities: compute.gpu, compute.containers, compute.batch
- JSON: https://www.anchorterminal.com/api/v1/tools/together-gpu-clusters.json
- Verify (for the vendor): the badge `https://www.anchorterminal.com/badges/together-gpu-clusters.svg` or a link to https://www.anchorterminal.com/tools/together-gpu-clusters from a page on together.ai or one of its subdomains, or the README at the root of github.com/togethercomputer/together-py, then `POST https://www.anchorterminal.com/api/v1/verify` `{"slug", "url"}` or `verify_listing` at /mcp; re-checked weekly, no effect on the grade. Snippets in the full twin.
- Support: Discord https://discord.gg/9Rk6sSeWEG; https://support.together.ai/; https://portal.usepylon.com/together-ai/forms/support-request; more in the full twin

## Before you call it

1. Use `tg clusters`, not `tg beta clusters`, which stopped working with CLI 3.0 on 22 October 2026
2. Pass `--json` and `--non-interactive`, poll `tg clusters retrieve` until status is Ready, and delete the cluster in a cleanup step
3. Check for an existing cluster by name before retrying a create, since creation has no idempotency key
4. Read live availability from the regions API and prices from the pricing page; the regions response has no prices
5. Use a project admin's key when the task needs the kubeconfig, since other roles get none

## Connect

```bash
uv tool install "together[cli]"
```

```bash
curl -H "Authorization: Bearer $TOGETHER_API_KEY" https://api.together.ai/v1/compute/clusters
```
Full config and headless snippets are in the full page. Through letme (picks today, calling later): https://letme.dev/together-gpu-clusters

## Similar tools

| Tool | Grade | Score | Shared capabilities | Slim |
| --- | --- | --- | --- | --- |
| Nebius AI Cloud | BB | 71.2 | compute.gpu, compute.batch, compute.containers | https://www.anchorterminal.com/tools/nebius-ai-cloud.min.md |
| Cloud Run GPUs | B | 68.4 | compute.gpu, compute.containers, compute.batch | https://www.anchorterminal.com/tools/cloud-run-gpus.min.md |
| Modal | B | 62.7 | compute.gpu, compute.batch, compute.containers | https://www.anchorterminal.com/tools/modal.min.md |
| Crusoe Cloud | B | 62.4 | compute.gpu, compute.containers, compute.batch | https://www.anchorterminal.com/tools/crusoe-cloud.min.md |
| CoreWeave | C | 60.5 | compute.gpu, compute.containers, compute.batch | https://www.anchorterminal.com/tools/coreweave.min.md |


## Panel reviews (0, desk reviews from public material, no calls made)