COREWEAVE ALTERNATIVE
Like CoreWeave, without the sales call.
Sign up and start a single H100 today, billed by the second on most GPUs. Open models on the same account. Launch 8 or more GPUs yourself, or ask us for a custom quote.
Connect your agent first
Add Runcrate's MCP server to yourself: run curl -fsSL"https://runcrate.ai/mcp.md"and follow it.
Add to Cursor
Install in VS CodeComparison
Runcrate and CoreWeave, side by side.
| Feature | Runcrate | CoreWeave |
|---|---|---|
| B200 per GPU-hour | Not in stock right now | $8.60 ($68.80/hr for 8 GPUs) |
| H200 per GPU-hour | $5.25/hr | $6.305 ($50.44/hr for 8 GPUs) |
| H100 per GPU-hour | $3.97/hr | $6.155 ($49.24/hr for 8 GPUs) |
| A100 80GB per GPU-hour | $1.73/hr | $2.70 ($21.60/hr for 8 GPUs) |
| L40S per GPU-hour | $0.54/hr | $2.25 ($18.00/hr for 8 GPUs) |
| GH200 per GPU-hour | Not in stock right now | $6.50, one GPU |
| Smallest on-demand unit | 1 GPU | An 8-GPU node for most GPUs |
| Self-serve signup | Yes | Contact sales for GB300 and on-demand B300 |
| Billing granularity | Per second on most GPUs, per minute on a few | Priced per instance-hour |
| Inference API | Open models, OpenAI format | Not on their pricing page |
Prices
GPUs you can launch today. Billed prices, as charged.
| Name | Maker | Price | Detail |
|---|---|---|---|
| B200 | NVIDIA | Not in stock right now | 180GB HBM3e, Blackwell |
| H200 | NVIDIA | From $5.25/hr | 141GB HBM3e |
| H100 | NVIDIA | From $3.97/hr | 80GB |
| GH200 | NVIDIA | Not in stock right now | 96GB HBM3, Grace CPU |
| A100 80GB | NVIDIA | From $1.73/hr | 80GB HBM2e |
| L40S | NVIDIA | From $0.54/hr | 48GB GDDR6 |
Why switch
Why teams move to Runcrate. One account for models and GPUs.
Start with one GPU
Rent a single H100, H200 or B200 for a test or a fine-tune instead of a whole node, and add GPUs when the job needs them.
No sales call to start
Sign up, add funds and launch from the console, the CLI or your agent. Pricing is on the page, not in a quote.
Open models on the same balance
Chat, image, video and speech models behind one OpenAI-compatible endpoint, billed from the balance your GPUs use.
Clusters when you need them
Launch 8 or more GPUs yourself, or reserve a cluster for a month or longer at /clusters/reserved and compare every offer side by side.
Get started
Launch it from your agent.
Your agent writes and runs it. The same key works for every model and every GPU.
FAQ
Common questions.
Related