[NOTE: ALL IMAGES IN THIS POST WERE CREATED USING GROK ON THE SAME DAY THIS POST WENT LIVE; IF YOU HEARD THAT NSFW IS DEAD, YOU HEARD WRONG]
Someone asked me this question on Reddit recently:
When I wrote this blog post, I explicitly told people NOT to subscribe to SuperGrok Heavy:
And I still stand by that.
This isn’t simply about the numbers, but we do need to put those numbers into context.
This is what I wrote about another video and how much content it took to actually create:
“A Long Goodbye to Grok” – which is featured in the blog post, “Should I Cancel My SuperGrok Subscription Because Grok is the Most Censored AI Platform?” – was made over the course of 5 days – from March 28, 2026 to April 1, 2026. So people need to really let that sink in: I generated 329 videos over 5 days, and that’s not counting failed attempts or video extensions.
The numbers cited by that user appear to be very similar to what the rate limits for SuperGrok used to be back in April – I think it matches up pretty well with what I documented at the time.
HOWEVER there is one important distinction: back in April failed generations didn’t count towards rate limits in the same way – if you take into consideration failed generation attempts, then while making “A Long Goodbye to Grok” I probably generated over 500 “videos.”
So even with the higher rate limits of current SuperGrok Heavy, at $100 a month you’re still receiving a service that’s inferior to what the $30 a month service offered back in April.
Back in April I made the argument that SuperGrok was the best value for AI image and video generation at $30 a month:
I would not make that argument now based on the current state of the service.
For the same reason, I can’t, in good faith, tell anyone that SuperGrok Heavy is worth it.
Any discussion about the value proposition of SuperGrok Heavy needs to confront the fact that the $99 a month price tag is temporary.
At the original price of $300 a month I don’t think it’s worth it for what most people are using it for as hobbyists.
Unless you’re REALLY hitting the limits for video every single day there are better ways to spend $300 on image and video generation.
I highly urge people to look at what other models are capable of:
Even at the $100 price range, I think other models and workflows become more viable.
The reason is because the raw numbers don’t give an accurate picture of how Grok functions in a real world use situation.
In terms of raw numbers SuperGrok Heavy sounds very competitive – because sure, 50 video generations a day sounds like a lot. Until you realize that you may easily lose half of those to moderation and there will be some things you simply can’t do.
It’s the same issue with image editing. Having 300 edits a day sounds like a lot, until you realize that a lot of edit requests will be moderated, burning a generation and producing nothing usable.
For a person focused mostly on image generation, $100 a month is insanely overpriced.
It’s not simply a matter of what the raw numbers are – it’s whether or not you’re going to be hitting the generation limits consistently every single day.
I just don’t think it’s realistic that a normal hobbyist using Grok is going to be generating thousands of images a month. In the 6 months that I’ve been a SuperGrok subscriber, I’ve rarely hit the image generation limits on Grok’s Speed setting – which can probably still generate at least 100 images a day. I’ve never even come close to generating 3,000 images a month.
So, I can’t envision a realistic situation in which an average user would need to perform close to 9,000 image edits every month.
Knowing that, there is no way I would pay $100 a month to use Grok for image generation. There are too many good and cheap alternatives that will produce more explicit content. Z-Image Turbo is a great example; it’s uncensored, it produces high quality images in multiple aspect ratios with good prompt adherence, and the cost per generation is $0.01 per image. If I need to edit something, Qwen and Seedream only cost $0.03 per edit and allow the use of multiple reference images. Seedream even allows me to upscale images up to 4k quality.
For video generation, the discussion is slightly different because video generation is significantly more expensive than image generation. But, even for video generation it’s a hard sell because there are a lot of other tools that can get the job done with a budget under $100.
And I’ll give some examples with real price tags on them so people can get a sense of how to assign a dollar value to these kinds of things.
This recent video was made in May 2026 using a hybrid workflow:
Half of it was made in Grok and half of it was made using Qwen 2.0 Edit and Wan 2.6 Spicy.
The image edits using Qwen 2.0 cost approximately $0.50.
There were 5 videos generated that cost approximately $5.90
So the total cost was approximately $6.40 for about a minute and a half worth of video.
The finished video is 3 minutes long.
This is another video that was produced in May 2026 and also used a hybrid workflow:
That video was mostly made in Grok, but there’s a 20 second segment [the park bench segment] that was created using Wan 2.6 Spicy and Wan 2.5 Extend. It probably costs about $5 to make the videos and another $3 to do the image edits with Qwen and Seedream.
So that was my video output in May – and while I did spend money outside of Grok, I’m well under $100. If you do the math, $30 + $6.40 + $8 = $44.40
And even if I had spent $100, the more explicit content in both videos simply couldn’t be made in Grok.
Even for my most ambitious project, Retro Rabbits, I don’t think I hit the $100 mark:
The videos generated outside of Grok cost about $22.50
The image edits were numerous; probably about 400 at a cost of around $0.36 per edit – which you can round up to around $15.50
So SuperGrok for $30 + 22.50 for Seedance videos + $15.50 for Seedream edits = $68, roughly
Which is a lot of money – but when you consider that I made a 15 minute short film for that budget, it’s not bad.
And again, the reason I HAD to use resources other than Grok was because, at the time, Grok did not allow users to create videos featuring nudity using uploaded images, even if the uploaded image had been made in Grok. So the only way to put the 2 main characters into a scene together [because they existed as different base images] was to use a tool other than Grok. Paying $100 for SuperGrok Heavy was not going to fix that problem.
So while I’m not arguing that SuperGrok Heavy isn’t the clear winner in terms of pure numbers, I think the discussion about value propositions is more nuanced.
I still like Grok and I intend to keep using it because I think it produces very good looking and believable images and video. But I am not going to be persuaded to pay $100 a month [or $300 a month], becasue at that point I am throwing money at a moderation problem that cannot be solved with money.
When Grok fails due to either moderation or rate limits, I am comfortable spending money on other models to produce the content that I want to produce. That’s what I recommend that everybody else do too. Embrace the idea of a hybrid workflow and don’t be afraid to try different things. Don’t be conned into paying more money for Grok and thinking that will solve all of your problems.
Leave a Reply