Menu

Ship ML to Prod, instantly. ⚡

Scale your machine learning inference and training on serverless GPUs.

How to Deploy on Banana

Select your deployment method.

Git-based deployment for custom Docker images. (setup required)

Models that deploy in seconds, no setup or source code required.

Click a model to deploy on 🍌

Enjoy 1 hour of free hosting on us 🍌

Only pay for the resources you use. That's the power of Banana.

	Usage Pricing	Enterprise Plan

	$.00051992	10-40%
	per second (A100 40 GB)	off usage pricing
A100 40 GB	$.00051992 /second	For companies spending $1000+ /month
A100 20 GB		$0.000330 / second
Egress Cost	none	none
Number of Seats	Unlimited	Unlimited
ML Model Size	up to 16 GB	16+ GB
Instant Deploys (Templates)	100	Unlimited
Custom Images	50	Unlimited
Secrets (build args)	Unlimited	Unlimited
Log Retention	7 days	7+ days
RAM (A100 40GB)	256 GB	256 GB
GPU RAM (A100 40GB)	40 GB	40 GB
vCPUs (A100 40GB)	16	16
GPU Concurrency	10	25+
CPU Concurrency	10	25+
Spike Tolerance	up to 10 replicas	25+ replicas
p99 Cold BootP99 is the 99th cold boot percentile. This means 99% of requests will be faster than the given cold boot number.	~50 sec	~10 sec
p50 Cold Boot	14.18 sec	14.18 sec
p99 System LatencyP99 latency is the 99th latency percentile. This means 99% of requests will be faster than the given latency number.	~13 sec	~13 sec
p50 System Latency	0.63 sec	0.63 sec
Network Payload	up to 50MB	up to 50MB
Support	Community Discord + Support Inbox	Private Slack channel w/ ~12hr SLA
	Deploy Now	Contact Us

Shout us out on Twitter to be featured here!