Notebooks
Modal Notebooks combine an intuitive interface with near-instant cold starts. This gives you the flexibility to run everything from lightweight experiments to large-scale multi-GPU training jobs.
Start a notebook running ML on high-performance GPUs in under 5 seconds.
Run code, see results, and refine ideas in real time, with the responsiveness needed for research and exploration.
Switch from CPUs to up to 8 Nvidia H100s or B200s in just one click.
Compute
Nvidia H100
80 GB @ $3.95/h
Nvidia T4
16 GB @ $0.59/h
Nvidia A10
24 GB @ $1.10/h
Nvidia L40S
48 GB @ $1.95/h
Use Modal's global distributed file system (Volumes) to work with massive datasets.
Access the latest GPUs without reservations. Pay only for what you use, down to the second.
Notebooks can call Modal Functions and share Volumes and Secrets with all of your Modal Apps.
'%3e%3cpath%20id='Vector'%20d='M14.4436%207.33336H11.3174C11.1778%204.09692%2010.0249%200.888916%207.99913%200.888916C5.97335%200.888916%204.82047%204.09692%204.68091%207.33336H1.55469V8.66669H4.68091C4.82047%2011.9031%205.97335%2015.1111%207.99913%2015.1111C10.0249%2015.1111%2011.1778%2011.9031%2011.3174%208.66669H14.4436V7.33336ZM7.99913%202.22225C8.7618%202.22225%209.84713%204.19203%209.98402%207.33336H6.01424C6.15113%204.19203%207.23647%202.22225%207.99913%202.22225ZM7.99913%2013.7778C7.23647%2013.7778%206.15113%2011.808%206.01424%208.66669H9.98491C9.84802%2011.808%208.76269%2013.7778%208.00002%2013.7778H7.99913Z'%20fill='%23AAADA8'/%3e%3cpath%20id='Vector_2'%20d='M8.00174%2015.1111C4.08085%2015.1111%200.890625%2011.9209%200.890625%208.00003C0.890625%204.07914%204.08085%200.888916%208.00174%200.888916C11.9226%200.888916%2015.1128%204.07914%2015.1128%208.00003C15.1128%2011.9209%2011.9226%2015.1111%208.00174%2015.1111ZM8.00174%202.22225C4.81596%202.22225%202.22396%204.81425%202.22396%208.00003C2.22396%2011.1858%204.81596%2013.7778%208.00174%2013.7778C11.1875%2013.7778%2013.7795%2011.1858%2013.7795%208.00003C13.7795%204.81425%2011.1875%202.22225%208.00174%202.22225Z'%20fill='%23AAADA8'/%3e%3c/g%3e%3cdefs%3e%3cclipPath%20id='clip0_0_6'%3e%3crect%20width='16'%20height='16'%20fill='white'/%3e%3c/clipPath%3e%3c/defs%3e%3c/svg%3e){kind=small}
ShareCreated by
gongy 2 hours agoLast edited by
gongy 2 hours ago akshat 10 hours agoViewers
sonad 5 min ago akshat 10 hours agoMultiple cursors, live edits, and shared context like in Google Docs—no more emailing .ipynb files.
Never waste money on a kernel you forgot to close down again, or configure if you need control.
Pyright type checking, rich outputs for visualizations, and AI-powered edit suggestions help you move fast.
Only pay for what you use, by the CPU cycle.
CPU
Physical core (2 vCPU equivalent)
$0.00003942 / core / sec
*minimum of 0.125 cores per container
Memory
$0.00000667 / GiB / sec
“Modal Notebooks have been incredible for sharing ML research across Suno. Engineers and designers can start playing with cutting-edge models minutes after they're finished training.”
“Other platforms like Colab didn't have the capacity we needed. We love how Modal Notebooks lets us scale up any amount of GPUs and compute.”
“Testing different GPUs on the fly has massively accelerated our workflow for profiling models. Tasks that used to take days now take minutes.”
Test and promote inference endpoints to evaluate quality, latency, throughput, and cost.
TypeSafe AI launched Jev, a new class of AI model built for automation, and scaled to trillions of tokens over a weekend on Modal.
Multi-node GPU clusters with RDMA, gang scheduled from Modal's shared capacity pool and billed by the second, behind a single decorator.
$30 / month free compute
84 ms