abacus.ai Reviews 177

My overall user experience has been great! I have used abacus.ai for brain storming and completed awesome projects using the platform.

Great experience! Both for newbies and for experts, Abacus.ai grows with your needs. I've been using it for the last year and the ChatLLM always gets you to the best model you need (you'll be always able to change it manually if you believe another one coul be cheaper or better). Other tools are easy and usable out of the box. Credit system is very clear and affordable (even if with an intensive use i only needed to top credit once!). Abolutely to try!

Very good answers. Love the quality.

They make new products each term.

Very, very good. I'm impressed.

It's an amazing tool as an aid in educational activities

Not an advanced user, but tapping into the chat and agent functionalities daily. Very pleased.

I've been using Abacus AI for over six months, and the biggest game-changer for me is the smart routing — the system automatically picks the best LLM for each task. I don't have to think about which model to use, it just works.

Before this, I was juggling multiple subscriptions. Now everything is in one place, and the quality hasn't dropped — if anything, it's more consistent because the right model handles the right job.

I'd recommend it to anyone who wants to stop paying for 3–4 separate AI tools and just get things done. It's not about being cheap — it's about being smart with your resources.

It is best from AI platforms.

The first package gives like 14K credit (whatever that is) and It cannot finish one agent run before it maxed out.
The credit are tooooooooooooooooo little for one incomplete agent output.
And the agent is not optimized for any tasks, the free version of claude opus 4.6 gives much better results.
This is a streight forward scam

I’m not satisfied with this service. In one day, I spent around €100, mostly through Deep Agent, without getting a genuinely working result. I gave very clear and logical instructions for specific bug fixes in an app, but the agent repeatedly claimed the tasks were completed, and when I checked, the changes were either missing or not working as described.

The biggest problem was that a lot of credits were used in long cycles without real progress. Instead of making targeted fixes, the agent often went in the wrong direction, made irrelevant changes, or stopped without delivering a finished result. That makes this service very poor value for money.

I would like this case to be reviewed seriously by their team, including the credit usage, the actual output, and whether the product is really performing as promised. In its current state, I would not recommend this service to people working on real projects and paying for reliability.

I used Deep Agent to create an app. Sheer torture. Not sure if it's the agent or me just being new to creating an app. Then, I needed a couple of folders that I couldn't access because It was locked in the Abacus system. Support, or what they call support just said, Just start over. They apparently couldn't be bothered. Another time after that I needed help and I was totally ignored. I'm hearing this is common with them. I'll probably bail.

The limit is a joke. It ran out in less than 24 hours! This is easily the worst service I’ve ever used. Super let down

Completely useless company. After paying an extra $10 to upgrade to the pro plan, for access to DeepAgent, it proceeded to waste all 18k credits in less than 40 minutes. Their support just stops responding if you try to get a refund, and the responses i did get were copy and paste. I've filed a fraud dispute with my bank

Awful, it's charged me for creating an agent, in fact, there is no agent (cause it don't debug it) and no results. Waiste of money

The platform does give you access to several models, but those that are worth anything when it comes to generating code burn through credits at an insane rate. Analyzing a small project (10-12 files under 400 lines each) and creating a summarization about half a page in length consumed almost 25% of the monthly credit allotment. Troubleshooting can be very cyclic and produce almost no results while burning through 100-500+ credits each attempt. The IDE itself is dysfunctional, and the agentic capabilities almost never work - code updates and insertions require a manual copy/paste. There are other options that are slightly more expensive, but provide a much greater degree of capability and reliability. I've canceled my subscription after 3 months, it's just not worth it in my eyes.

The deep agent was the reason I signed up. It basically stopped working after I added more prompts. Then when I tried to start a new topic, after adding only a few sentences of what I wanted I had to end the session. When I returned nothing was saved and it counted as one of the three deep agent tasks you get each month with your subscription.

Got no support from the company and would not recommend this product and give it a one star rating.

I used Abacus.ai for over a year now and just cancelled because of the price and token usage of the coding agent. I had a small fix for one of my projects, and it ended up costing me about 10,000 credits because it would just loop and make the same mistakes. The most use I got out of this product is the projects feature but most platforms have this feature as well.

Website is shit, use your tools to make it look good first,
if you can't do that then you are contradicting your business model.