I can never get a gpu but I'm paying…
I can never get a gpu but I'm paying each month
Monthly average ratings based on 50 reviews
Showing 50 of 50 reviews
I can never get a gpu but I'm paying each month
Limited resources, poor service, and unreliable information during deployment. Virtually unusable. A complete waste of money!
crypto miners still running the site, no affordable gpu available for rent because of that.
The resources availability needs to be improved.
freaking expensive, almost like a scam. I sent them a few emails but no reply.
So bad, don’t recommend this
Very intuitive to use, and both pod and serverless endpoints work like a charm. I've been with them for almost 4 years, they have had very few outages so far, unlike many other cloud providers. Thanks for making such a great platform!
I rented a server with an RTX PRO 6000 on Community Cloud. After a lot of investigation into why my LLM was running four times slower (in terms of tokens per second), I discovered that the GPU had its power limit set to half, which restricted it to only a quarter of its potential clock speed (600 MHz). I gathered evidence using various commands and other data. After submitting a support ticket, they refused to issue any refund for the Community Cloud service, stating: "No refunds are possible under any circumstances for Community Clouds." Well, both my money and my time were wasted and through no fault of my own. RunPod does not protect its customers.
GPUs were already scarce for rental which made the experience mediocre, but MTL-1 has been down for days and EU-1 which went down almost for as long is barely functioning properly based on user reports. You can't even boot up a CPU pod to download your data. What's worse is that they will not explain why the data centers are down, despite this being the longest outage they've ever had. They've also been jacking up GPU rates (understandable) without advanced notice (not cool). Runpod is building a reputation of lacking transparency and lacking data security - you will not find peace of mind here. Negative reviews are their just desserts for no public communication for the latest outages; but given how GPUs aren't lacking in users, maybe downsizing their potential customer base would be doing them a favor.
Except for the price everything is excellent :b
Worked as advertised
Wonderful application.
Runpod APIを自作スクリプトで呼び出すことでPod確保がしやすくなった。 Network VolumeもAPIで呼び出せるのでPod起動せずとも内部の変更ができる点もよい。
This is my go to for cloud gpu! I have a decent local AI setup. But when I need an ultra fast AI engine for an AI system I'm building I use Runpod. The prices are excellent and ease of use are excellent!
RunPod never takes accountability for non-functional pods... Broken CUDA drivers, intentionally limited performance, no way to see how reliable a pod is before being assigned to it... 50% of the money spent on runpod so far has been just waiting to see if the pod is even functional, I would never recommend this service to anyone EDIT: Just to be clear, this review was made after multiple different tickets to support about different pods... all of them resolving to "not our problem"
I spent hundreds of dollars operating pods that got me nowhere and wasted many months of my time having to configure and reconfigure and fix issues. Customer service and their way of addressing issues is unprofessional and basically say issues are the past even though its recent, and they refuse to address complaints. Not professional. I do not recommend using this company.
Belle bande d'escrocs.... incompatibilité à tous les étages, gpu qui te lâche sans explication... le seul truc qui marche c'est la facturation ... ah là dessus ils sont bons et c'est au centième de seconde ... Je paie le stockage, mais aucun GPU n’est jamais libre : j’attends depuis des heures. Interface incompréhensible, erreurs absurdes, disque inaccessible. Ce n’est plus un service, c’est du vol. À fuir. bande de voleurs.
RunPod's support is the weakest I've dealt with. When something goes wrong, they can't find the cause. You get suggestions to change your setup or spend more, with no idea why. You end up doing the investigation yourself, and when it's finally escalated, they call it your fault for following their advice. The ticket ran for weeks and was then closed with "no further discussion." This isn't the first time it's happened to us. If you run anything in production, don't count on their support.
Terrible experience over all. Could not get anything to work. Now stop sending me emails
Todo perfecto!
many of the requests get stuck because there is no GPU capacity.
Fairly easy to set up with an agent, thanks to which I was able to run some autonomous workflows without me needing to do anything manually. The only thing is that the pricing was ambiguous at first - I had to understand that the $0.79 5090 and 0.35 one have different capabilities and the more expensive one is the one I should be going for.
**RunPod Needs a GPU Reservation System — Not Just a Waiting List** I can understand RunPod's business model. Of course, the company wants its GPUs to be running 24/7 rather than sitting idle. But honestly, I would not recommend RunPod for serious projects if you have to wait for a GPU to become available. Switching to another GPU instance is not always a simple solution. You risk losing your data, configuration, installed dependencies, and hours of work. Data loss may not happen every time, but migration can be complicated and risky, especially when you are working on serious projects. There is a very simple solution that could benefit both RunPod and its customers: **a GPU reservation system.** Imagine a reservation calendar displayed next to the available GPUs. Users could see which time slots are occupied and which are available, then reserve a GPU for a specific period. The payment could be charged upfront, or automatically two hours before the reservation begins. Once the reservation is confirmed, the customer knows that the GPU will be available at the agreed time. This would give customers predictability, eliminate unnecessary waiting and stress, and make it easier to plan serious development work. At the same time, RunPod would have better visibility into future demand and could monetize its GPU capacity more efficiently. It is a straightforward idea. Instead of forcing customers to constantly check availability or migrate their projects between instances, why not let them book the computing resources they need? **RunPod, this is a win-win opportunity.** You could improve the customer experience, build customer loyalty, and make better use of your infrastructure. Perhaps someone at RunPod will hear this suggestion and turn it into a real feature.
Refund me as many times I don't find available GPUs.
It just works. Web interface is clean, GPU selection is very good, prices are fair. Had no issues at all so far. I spin up a runpod for ComfyUI diffusion model images inference once in a while and with global persistent storage connected. Super convenient and nice user experience. The only thing that I was missing is a queue Runpod feature on the selected GPU when it is not available. Because my favored GPUs are almost never available and I usually have to wait a few minutes or up to about half an hour for it to become available and me catching it. Hence a queue feature would be great. Or obviously increasing GPU capacity for low price ones.
Sent me this spam email survey I never agreed to
I had lots of issues with software taking forever to load as well as there being no available graphics cards for a lot of the times I wanted to use them. I wasted a lot of money because it would take so long for the software to download and also had issues with memory exceeding the limit and crashing the pod
Whne you are a beginner, it's to much hard. lucky for me that Gemini help me a lot.
Overpriced and cumbersome to use. The UI is blocky and not intuitive. With that said, It’s a decent alternative while commercial GPU and VRAM are expensive, but you're always better off buying instead of renting.
Overall it works as I'd expect, only issue is GPU availability, but I also get that it would be more expensive if they had a ton of idle GPUs just sitting there all the time.
Perfect systems
Lack of pod availability , it can be very difficult to get capacity if you don’t keep your pod running 24/7 edit: i am typically running RTX 4090, and that is a BS reply. almost everyday i am constantly retrying to deploy my pod
Machine are not available, when I required. I getting higher or lower version
I use RTX 6000. It runs very slow. Almost all videos I generate have problems of strange language in addition to the English, thus spoiling MOST of my videos.
A plataforma em si é muito boa, porém não vou negar que a falta de placas disponíveis e atreladas a volumes por região deixa um pouco a desejar. Muitas das vezes vou iniciar um Pod, ou uma requisição Serveless e tenho q procurar regiões com GPUs disponíveis, e com isso, pra suprir meus modelos de operação tenho q ter diversos volumes e pagar diferentes valores por eles (uma coisa que o armazenamento Global era pra suprir, mas convenhamos que ele é lento de mais para isso. Não compensa ir de aproximadamente 8-9 minutos de tempo de produção por solicitação para 25-30 minutos só por causa do volume).
caro, lento e sem graça. Mesmo alugando uma maquina forte as gerações demoraram muito, muitos bugs no comfyUI.. enfim, não recomendo.
it is totally functional and usable. sometimes i feel it is costly compare with other available sites. if the GPU and storage cost could be slightly reduced, then it will grab more audiance...cheers!
Майже завжди доступні відеокарти, по добрій ціні, сервіс супер, сапорт відповідають швидко👍
It’s definitely a great service when you don’t have the compute power needed for projects!!
Expensive for what you get. Cheaper GPU like rtx4090 also has limited availability.
Easier to figure out more capable and more easy to configure the payment than Floyo, which was the cloud service I tried right before this. I like that I can use any workflow and any node setup I want and that I can use anything from Civit.ai Thank you! You guys made it actually affordable to run these big powerful workflows that other creators put out that I wanted to try.
Very good company. Great servers and very cheap. When the old server is busy, do a quick migration between pods
It generally worked well and was cost effective
Would be great platform to run CI. But because docker-in-docker doesn't work I cannot run 90% of my burstable workloads.
Well documentation, easy to start to deploy right way and api web socket connection is amazing. Will provide your services very soon on instagram. Thank you much for your work.
es ist kaum möglich einen prozess zum laufen zu bringen, da kedesmal alle grafikarten ausgebucht sind. Abslites nogo als kleiner unternehemer einfach nciht brauchbar. Nur wer viel geld hat und sich eine dauermiete leisten kann ist im vorteil.
Have a question about Runpod? Ask our AI assistant powered by customer reviews.
Reviews are aggregated from Trustpilot. We do not edit or remove reviews.