Anyscale Reviews & Product Details

Top-Rated Alternatives

Databricks Data Intelligence Platform

View All Alternatives

Anyscale Integrations

(1)

Integration information sourced from real user reviews.

G2 reviews are authentic and verified.

Here's how.

RA

Rakshit A.

AI Application Engineer

Mid-Market (51-1000 emp.)

11/19/2025

"Great tool for scaling AI workloads"

4.5/5

What do you like best about Anyscale?

What impresses me most is how it handles the heavy lifting for Ray. I can develop my AI application code right on my laptop and then deploy it to a large cluster without having to rewrite anything or wrestle with complex infrastructure setups. This effectively bridges the gap between code that only "works on my machine" and a real production environment, which is particularly useful when scaling LLM workloads and managing distributed training. In the end, it saves me a considerable amount of time on DevOps tasks. Review collected by and hosted on G2.com.

What do you dislike about Anyscale?

The pricing structure can feel somewhat unclear, making it difficult at times to anticipate your final monthly bill. This is especially noticeable when compared to the more straightforward cost management you get with handling raw EC2 instances on your own. Review collected by and hosted on G2.com.

What problems is Anyscale solving and how is that benefiting you?

I use Anyscale mainly to overcome the infrastructure challenges of scaling Python machine learning code from my local laptop to a large distributed cluster. My team operates a substantial Retrieval-Augmented Generation (RAG) pipeline, which includes OCR processing and embedding generation for millions of PDF files. Previously, running this workload on a single large EC2 instance would take weeks, and managing AWS Batch jobs involved a lot of boilerplate and ongoing DevOps work. With Anyscale, we were able to wrap our existing Python functions with Ray decorators, enabling the platform to automatically spin up a cluster of over 50 spot instances, process 2TB of data in less than four hours, and then scale back down to zero. This approach has reduced our compute costs by about 60% by taking advantage of spot instances without the need for manual fault-tolerance solutions, and it has allowed my data scientists to independently run large-scale experiments without waiting for DevOps to provision resources. Review collected by and hosted on G2.com.

SM

Subrat M.

Senior DevOps Engineer

Mid-Market (51-1000 emp.)

8/25/2025

"Scalable and reliable platform for AI workloads"

4.5/5

What do you like best about Anyscale?

Anyscale simplifies the process of moving AI and ML workloads from development to production. Since it is built on Ray, it enables scalability without requiring major code changes. Review collected by and hosted on G2.com.

What do you dislike about Anyscale?

The platform has a noticeable learning curve, particularly for teams unfamiliar with Ray concepts. Pricing is not always transparent, which makes cost planning more challenging. Review collected by and hosted on G2.com.

AR

Akanksha R.

Web Development Intern

Computer Software

Mid-Market (51-1000 emp.)

8/19/2025

"Good"

4/5

What do you like best about Anyscale?

The Anyscale platform was essential as it fully managed, production-ready version of Ray, offering a simplified and integrated developer experience and it made easy to build and it has good scalability. Review collected by and hosted on G2.com.

What do you dislike about Anyscale?

About the disadvantage is during building there is little trouble when debugging trouble. Review collected by and hosted on G2.com.