Google introduces Colab Pro w/ faster GPUs, more memory, and longer runtimes

Google Colab is a useful tool for data scientists and AI researchers sharing work online. The company this week quietly introduced a paid “Colab Pro” tier with three benefits.

Within a Colaboratory’s interactive environment, users can write and execute Python on the web with zero configuration, free GPU access, and easy sharing through the exact same Google Drive interface. It’s particular useful for students in educational settings and others learning about ML.

Colab Pro is an upgrade that provides three primary benefits for $9.99/month:

• Faster GPUs: “Priority access to faster GPUs and TPUs means you spend less time waiting while code is running.” Instead of free K80s, subscribers get access to the Nvidia Tesla T4 and P100, as well as being prioritized for Tensor Processing Units. Google notes that there are still usage limits with Colab Pro.
• Longer runtimes: “Longer running notebooks and fewer idle timeouts mean you disconnect less often.” Instead of 12 hours, Colab Pro lets notebooks “stay connected for up to 24 hours, and idle timeouts are relatively lenient.”
• More memory: “More RAM means better performance, and less running out of memory.” Notebooks will feature a “high-memory VM” preference that “generally have double the memory of standard Colab VMs, and twice as many CPUs.”

Colab Pro does not guarantee resources to “maintain the flexibility to adjust usage limits and hardware availability on the fly.” That said, Google says intended interactive computing users will “experience few if any usage limits.”

The paid offering was quietly introduced this week with a retweet from Google AI head Jeff Dean bringing it to light on Saturday. The free tier remains unchanged, with the subscription currently only available in the US. Google made a demo notebook with Colab Pro to explore the benefits.