Free GPU notebooks on Paperspace Gradient (now part of DigitalOcean) get you reliable access to an NVIDIA Quadro M4000 (8GB VRAM), plus 5GB persistent storage and 6-hour sessions you can restart endlessly. If you’re searching for DigitalOcean free credits but you mostly need GPU time for experiments, this is one of the more practical “$0” options out there.
ML engineers running quick benchmarks, founders validating a model before paying for real infrastructure, and students trying to finish a class project without a GPU bill tend to get the most value here. Just know what you’re signing up for. The good stuff is real, but it comes with tradeoffs.
This guide breaks down eligibility, the exact signup flow, the service limits that matter (public notebooks, session shutdowns, GPU availability), and a few ways to stretch the free tier further.
Program at a Glance
| Provider | DigitalOcean (Paperspace Gradient) |
| Credit Amount | Free GPU notebooks (M4000 8GB reliably) |
| Duration | 6-hour sessions, unlimited restarts |
| Eligibility | Anyone who can verify with a credit card |
| Credit Card Required? | Yes, required for verification |
| Difficulty | Easy; standard signup + notebook creation |
| Best For | Jupyter experiments, prototypes, light training |
| Official Page | DigitalOcean Program Page |
What You Actually Get
Paperspace Gradient’s free tier gives you GPU-powered Jupyter notebooks you run in the browser. The “reliably available” free GPU is the NVIDIA Quadro M4000 with 8GB of VRAM (Maxwell architecture), and you also get a Free-CPU option (C4). Paperspace advertises free access to higher-end Ampere GPUs like the A4000/A5000/A6000, but those are heavily constrained by availability and often grayed out when you try to start a notebook. You also get 5GB of persistent storage and a fixed 6-hour auto-shutdown per session, with unlimited restarts and no weekly GPU hour cap.
In real-world terms, this is enough to iterate on notebooks, run smaller fine-tunes, test data pipelines, and do reproducible experiments with saved checkpoints. It’s not a free A100 box. Honestly, it’s better to think “free M4000 notebook platform, sometimes better hardware shows up” and plan your work accordingly.
Who Qualifies (and Who Doesn’t)
Gradient’s free tier is broadly available, but it is not “no-strings-attached.” You need an account and you must verify it with a valid credit card, even if you only plan to use free machines.
- You need a valid credit card for verification, and Paperspace processes it through Stripe.
- Visa, MasterCard, Amex, and Discover are accepted for the required verification step.
- The free tier is designed for Gradient Notebooks, not Deployments or Workflows.
- You have to be comfortable working in public notebooks on the free plan.
If you can’t provide a credit card, you won’t be able to activate the free tier. And if you need private notebooks, terminal access, or production features, the free tier won’t qualify you for those.
How to Sign Up
Signup usually takes about 10 minutes if you have a card ready.
- Go to paperspace.com and click Sign Up.
- Create an account with email, or sign in via Google or GitHub.
- Enter a valid credit card for verification (required for the free tier; Stripe supports Visa, MasterCard, Amex, and Discover).
- From the dashboard, navigate to Gradient > Notebooks.
- Click Create Notebook, then pick a runtime template like PyTorch, TensorFlow, or Start from Scratch.
- Under Machine, select a Free GPU/CPU option (Free-GPU M4000, Free-CPU C4, or check whether Free-A4000/A5000/A6000 show as available).
- Click Start Notebook to launch your in-browser Jupyter environment.
After you start the notebook, it launches directly in your browser. If a free GPU is unavailable, you may be queued or blocked from starting until capacity opens up, especially for the rarer GPUs.
What the Credits Cover
This “free tier” isn’t a dollar credit balance you spend down. It’s access to specific free machine types inside Gradient Notebooks, plus a small amount of persistent storage that stays across restarts. The practical scope is: run Jupyter notebooks for ML experimentation, and save artifacts under the persistent storage directory so you don’t lose work between sessions.
| Service / Feature | What It Does | Included? |
|---|---|---|
| Gradient Notebooks | In-browser Jupyter notebooks for ML work. | ✓ |
| Free GPU machines | Run notebooks on shared GPUs like M4000 (8GB). | Partial |
| Persistent storage (5GB) | Storage that persists between notebook restarts. | ✓ |
| Deployments & Workflows | Production deployment and workflow orchestration features. | ✗ |
Notable exclusions: the free tier does not include Deployments or Workflows, and you don’t get terminal/shell access. Also, the A100 is not free (it requires a paid plan plus hourly pricing, around $3/hour).
Limitations to Know About
Every free GPU program has catches. With Gradient’s free tier, the limits are mostly about session time, privacy, and shared capacity.
- Each session auto-shuts down after 6 hours, but you can restart immediately and keep going.
- The free tier allows only 1 concurrent notebook running at a time.
- You get 5GB of persistent storage, and overages cost about $0.29/GB/month.
- Free tier notebooks are public, so you should not store API keys or proprietary code there.
- There’s no terminal access on the free tier, which can be annoying for debugging.
- Free GPU availability is not guaranteed, and you may be queued during peak demand.
- Free machines can be used for Notebooks only, not for Gradient Deployments or Workflows.
- You’re limited to up to 5 Gradient projects on the free tier.
When credits (capacity) run out, you don’t “lose” an account balance because there isn’t one. What happens is simpler: you just can’t start the machine you want until it’s available, or you move to a paid plan and paid on-demand GPUs. The 6-hour shutdown ends your current session, but unlimited restarts mean you can resume work as long as you saved what matters to persistent storage.
Have Unused DigitalOcean Credits?
A lot of teams end up with DigitalOcean credits they can’t spend in time (especially when credits come from startup programs or internal cloud budgets). Those credits typically have expiration clocks, and watching them hit zero is painful. If you have surplus DigitalOcean credits you won’t use, AI Credit Mart lets you list them so someone else can put them to work at a discount. It’s a practical way to recover value instead of letting credits expire.
Need More DigitalOcean Credits?
If you outgrow the free M4000 notebooks, the next step is usually paid GPUs or a paid plan. You don’t always have to pay sticker price, though. AI Credit Mart lists discounted DigitalOcean credits from companies with surplus allocations, and discounts commonly land in the 30–70% range. That can stretch your runway, especially if you’re doing a few months of GPU-heavy work.
Tips for Getting the Most Out of Your Credits
- Keep secrets out of the notebook because free notebooks are public; pass credentials at runtime instead of hardcoding them.
- Save checkpoints to persistent storage (the /storage directory) so you can recover cleanly after the 6-hour shutdown.
- Check GPU availability during notebook creation since unavailable free GPUs show as grayed out, and try a different time or region if needed.
- Use the pre-built templates (PyTorch, TensorFlow, Hugging Face, Stable Diffusion, fast.ai) to avoid spending half your session on environment setup.
- Consider Google Colab if you specifically need a stronger free GPU than the M4000 and you can tolerate idle disconnects.
Frequently Asked Questions
They’re not a fixed dollar balance; the value is free access to shared notebook machines, most reliably the Free-GPU M4000 (8GB VRAM), plus 5GB persistent storage. In practice, you can run unlimited 6-hour notebook sessions with restarts, which is plenty for coursework, quick model experiments, and iterative prototyping. The “worth” depends on what you would otherwise pay for GPU time, but the big win is no weekly GPU hour cap. Treat any free Ampere GPU access as a bonus, not the baseline.
Yes. A valid credit card is required for account verification, even on the free tier.
For this program, the key limit is the 6-hour auto-shutdown per notebook session, and you can restart sessions without a weekly GPU-hour cap.
Yes. If you have DigitalOcean credits you won’t use before they expire, you can list them on AI Credit Mart and sell them at up to 70% of face value. Companies regularly list surplus credits from startup programs and enterprise agreements.
AI Credit Mart has discounted DigitalOcean credits available from companies with surplus allocations. Prices are typically 30-70% below retail.
Gradient’s free tier isn’t a spendable credit balance, so there’s nothing to “expire” like a coupon. What you will hit are usage limits: your notebook shuts down after 6 hours, and you may be unable to start a free GPU when capacity is tight. If you move to paid on-demand GPUs, that usage is billed hourly (for example, A100 instances are priced around $3/hour). So the practical outcome is: free access can be unavailable or time-limited, and paid usage is billed normally.
Some Ampere free instances exist but are frequently unavailable, and the A100 is not free (it requires a paid plan plus hourly charges).
Not really. Free tier notebooks are public, so you should not put API keys, credentials, or proprietary code in them.
Gradient’s free tier is legit for what it is: a dependable free M4000 notebook with persistent storage and no weekly GPU-hour cap. Use it to learn, prototype, and iterate fast, and if you ever end up with surplus DigitalOcean credits, you can sell them instead of letting them die on the vine.
Your AI credits are losing value every day
Join the marketplace and start trading unused credits today.