What you get, and what it costs
Numbers you need mid-decision should not be buried mid-paragraph across nine units, so every reference table lives on this one page. First pass: read the credits table and set your budget alerts before you touch anything billable. After that, come back whenever a decision has a dollar sign in it.
Figures as published in the source guide. Pricing, tiers, and hardware move — verify anything you are about to spend against before you spend it.
Subscription tiers at a glance
| Plan | Cost | Storage | Key AI access |
|---|---|---|---|
| Free | $0 | 15 GB | Limited Gemini, T4 GPU when available |
| AI Plus | $7.99/mo | 200 GB | Expanded Gemini usage, Veo 2 video generation |
| AI Pro | $19.99/mo — free for students | 5 TB | Gemini 3.1 Pro, Deep Research, L4 GPUs, MCP |
| AI Ultra | $249.99/mo | 30 TB | Highest quotas, $100/mo cloud credits, A100 GPUs |
Colab GPU access by tier
| GPU | VRAM | Best for |
|---|---|---|
| T4 — free tier | 15 GB | Prototyping, fine-tuning small models, most coursework |
| L4 — Pro ($11.99/mo) | 22.5 GB | Transformer fine-tuning, medium training runs |
| A100 40 GB — Pro+ | 40 GB | Large model training, LLM work, serious research |
| A100 80 GB — Pro+ | 80 GB | Very large models, multi-GPU experiments |
Must-have MCP servers for data scientists
| Server | Category | What it does |
|---|---|---|
| GitHub | Source control | Review PRs, track issues, analyse repo history |
| BigQuery | Database | Query schemas and reason over live data in natural language |
| Firecrawl | Web scraper | Scrape any URL into clean text for literature reviews |
| Hugging Face | AI hub | Browse models, datasets, and papers from the terminal |
| Notion | Knowledge base | Access team specs, roadmaps, and decision docs |
Choosing the right tool
| Task | Best tool |
|---|---|
| Chat-based research, writing, summarising documents | Gemini Advanced (browser) |
| Search your Drive files and email with AI | Gemini + @Drive / @Gmail |
| Automate file ops, run shell scripts, generate code in a terminal | gemini-cli |
| Interactive data analysis, model training, Python notebooks | Google Colab AI |
| Multi-agent coding, full codebase navigation | Google Antigravity (preview) |
Credits and guardrails
- The AI Pro subscription includes $10/month in Google Cloud credits — $120 across the year — on top of the $300 free trial every new GCP account receives.
- Activate the monthly credits at developers.google.com/profile after setting up the student offer.
- Credits are routed through Vertex AI. They do not cover direct Gemini API calls made through Google AI Studio; those accrue separate charges.
- Set budget alerts at 50% and 90% before running any GPU jobs, under Billing → Budgets & Alerts.
- Budget alerts send emails only — they do not stop services automatically. Spend caps took effect April 1, 2026.
- PhD students can apply for Research Credits — up to $1,000/year for coursework-adjacent projects — at edu.google.com/programs/credits/research with a short proposal.
- Vertex AI: managed ML training, model deployment, and AutoML — a production-ready Colab with persistent environments, guaranteed GPU access, and real API endpoints.
- Model Garden: one-click access to 200+ models including Gemini, Llama, Mistral, and Stable Diffusion.
- Document AI: structured data out of PDFs, scanned documents, and forms. Basic OCR covers roughly 200,000 pages on a $300 credit.
- GPU VMs: dedicated T4, L4, or A100 instances with no session timeouts. Spot VMs run 60–90% cheaper — pair them with checkpoint saves.
- BigQuery: 1 TB of free SQL queries per month plus 200+ public datasets, pullable into Colab with one line of Python.
- The student offer is currently US-only for new sign-ups, and auto-renews at $19.99/month unless cancelled before twelve months are up.