UPDATE: we’ve updated this post to include a new bike. See the current info here >
Zwift released its first mountain bikes in last week’s major update. The Canyon Lux, Scott Spark RC, and Specialized Epic S-Works join the “Zwift Mountain” to make a total of 4 MTB rigs available.
Scott Spark RC Specialized Epic S-Works Canyon Lux
But how do they actually perform in game? How do they compare to the established crop of road rigs, and Zwift’s new gravel bikes?
We ran extensive tests to find out. Here are some of the results.
Flat Pavement Performance

Here’s how each mountain bike performed over two laps of our Tempus Fugit test segment (300 watts steady, 75kg rider):
- Specialized Epic S-Works: 55:35
- Scott Spark RC: 55:35
- Canyon Lux: 55:36
- Zwift Mountain: 55:43
The Spark and Epic S-Works turned in the same fastest times, but the Lux is only 1s behind, which can be a Strava rounding error. We’ll give the win to all three.
By comparison, the fastest road bikes complete the test route in 50:25, while the slowest road bikes (Zwift Steel with 32mm carbon wheels) complete it in 51:40. So if you’re riding an all-pavement race, you definitely don’t want to be riding a mountain bike!
The fastest gravel bikes complete the test route in 53:47, which is what you would expect: the mountain bikes’ big knobbies and poor aero values make them slower than road or gravel rigs on pavement.
But how do the mountain bikes perform in other settings?
Climb Performance

Here’s how the mountain bikes performed up the big Alpe du Zwift climb:
- Scott Spark RC: 54:29
- Specialized Epic S-Works: 54:30
- Canyon Lux: 54:40
- Zwift Mountain: 54:50
The Spark RC and Epic S-Works are essentially tied after almost an hour of climbing, with the other MTB rigs well behind.
The fastest road bikes climb the Alpe in 48:39, while the slower (Zwift Steel with 32mm carbon wheels) complete it in 49:57. So the heavy mountain bikes are well behind, as they should be. Because gravity matters!
Jungle Performance

What about the dirt? This is where knobby tires should excel, and indeed they do. Here’s how the mountain bikes perform on one lap of the Jungle Circuit:
- Specialized Epic S-Works: 13:51
- Canyon Lux: 13:52
- Scott Spark RC: 13:52
- Zwift Mountain: 13:54
So the three new MTB rigs turned in nearly identical performances, with the stock Zwift Mountain bike a few seconds behind.
The faster road bikes complete a jungle lap in around 15 minutes flat, while the gravel bikes do it in 14:11. But our mountain bikes shred some gnar and lay the hurt on all those skinny-tired, rigid-forked rigs!
MTB Conclusions
Scott Spark RC Specialized Epic S-Works
The Scott Spark RC and Specialized Epic S-Works are clearly the top two bikes of the four MTB rigs. The three new mountain bikes have the same aero values, but the Spark RC and Epic S-Works are lighter, making them the winners. Which one should you choose? Whichever looks cooler, of course.
Interesting side note: Zwift retooled the Zwift Mountain bike with their latest release, reducing its performance in game.
Your Thoughts
Have you tried the new mountain bikes? What did you think? Share below!

Always great work. Appreciate your research!
Is the time of the Zwift Mountainbike on the Alpe right?
Whoops! Nope. Fixed.
Another Epic report from Eric. Interesting to note that they all compare virtually identically. I rode the Jungle on the Zwift MTB the other day and flew past a group of TT bikes then on the climb out towards swing bridge I caught the bloke in front on a normal road bike by some 1 min 20 seconds. We were both outputting nearly the same watts.
Will be interested to see where Zwift go with this gravel routes etc as the pure roadies and Segment busters are moaning like crazy at going a bit slower. 🙂
And the war against the hardtail wages on…
The fact that the Specialized Epic S-Works & the Scott Spark RC are the same performance is BS. Why does the Specialized cost 200,000 more drops for the same experience?
Apparently the Specialized brand’s notorious high cost is enforced virtually as well as IRL.
You are correct. Prices in Drops are directly derived from real-world retail prices. Zwift chose to do it that way across the board.
Would be useful to have level locking included in these reports … i.e. at which level we can actually buy these.
For the Jungle, there is very little difference, and the Zwift Mtn bike is right next to my Tron in the garage. Super easy to switch, and cost me no drops. Still wish the gravel bikes got their just due in the Jungle, as they should be faster on dirt roads.
I completely agree with the point on gravel bikes. I don’t understand how a mountain bike could be faster than a gravel bike on the jungle circuit. It’s packed down dirt and wood elements. Zwift should really recalibrate this. Make the mountain bike faster on the more technical mountain bike tracks and the gravel bike faster on the dirt and wood. I’m looking forward to buying the Aspero. It should make a difference.
This is an excellent point. I was contemplating buying one of the non-Zwift bikes. It makes sense, however, to put the focus on bike swapping time rather than performance around the jungle loop. So if I use the Tarmac Pro or the Z1 on the road, then either the Scott or Specialized will be next to the Tarmac, while the Zwift mountain bike would be next to the Z1. Basically the best bikes (Scott, Epic) save 20 seconds per 50 minute of climbing. 8 seconds per hour of flats. The tires are the same. Dirt increases the importance of weight… Read more »
Eric, can you benefit from draft while riding a MTB (or gravel bike)?
Yep.
Thanks for your swift reply Eric, and thanks for all your amazing help for new Zwifters like me (new since +- 2,5 months), they really made a difference!
Hi Eric,
I am now reading your articles since I started Zwift November last year.
I have to admit this is quite addictive and it can lead to overtraining 💁
I have been gathering all your data into an Excel file in order to compare all the times between road bikes, TT, gravel, MTB…
When you mentioned the fastest bike competed two laps Tempus Fugit in 50:25, which bike is it?
I do not find it in the Excel extracts you share in your articles.
Thanks
Thomas
Indeed another helpfull test!
What I would also be interested in; is the weight penalty for light riders the same on dirt compared to tarmac?
Would be another interesting simulation.
Eric, Any plans to conduct a speed test for a course like Road to Ruins where you’ve got a mix of tarmac and gravel/dirt and how over the course of that…well, course, one of the faster road setups would fare against faster gravel or MTB? This test would assume same bike for the entire course and of course the notion is pitting the cumulative speed advantages of the road bike on the tarmac vs. the speed advantages of gravel/mtb on the dirt, with the notion being….you don’t want to switch bikes mid-ride based on surface and take the switch time… Read more »
Hey @guest_35557: you inspired me! https://zwiftinsider.com/road-to-ruins-recommendation/
Thanks for taking the time to do that Eric! So cool…..and BTW…thanks for doing all that you do with your testing and comparisons….I absolutely rely on the information you provide to maximize my Zwift experiences, even though I’m a D class rider and by no means a serious competitor even in that class.
Eric, that was my comment above attributed to “David Cooper” Didn’t realize the comments had a “generic” user setup and that my work computer hadn’t remembered my previous login settings….before I pushed the post button. Doh. It’s definitely Friday.
It’s a weird bug in our comments plugin, @guest_35557 … every once in a while someone posts a comment and it posts as being from someone totally different who has previously commented. Annoying!
That was weird … was wondering whether I needed to change my posting handle now that there was more than one of me.
Eric. I’d be fascinated to see what you are going to ride in the next stage of the Tour of Watopia around the jungle.
I’m trying to decide if an MTB would be OK in a big group or whether a gravel bike would be better for the final sprint.
I’m going MTB, and riding in a head-up manner so I (hopefully) don’t get dropped on the wood/pavement portions!
I just noticed on the official site that they are recommending MTB, so I guess a lot of other people will be riding them. As you say, it is only the wooden/pavement sections where the gravel bikes have the edge.
Thanks!
This site is SO helpful in prepping for races!! Looking forward to today’s Classic in the Jungle!!
Excellent article, as always. It’d be nice to eventually see some mtb wheel upgrade options in the drop shop!
Hi…I’m wondering what percentage slower a mountain bike is compared to a road bike…roughly (I accept that this will differ depending on which bike is used etc).
For example, if a road bike rider is putting out 2.5w/kg on a flat tarmac course, what would the same weight/height rider need to push to keep up on a mountain bike…50% more (3.8 ish), 25% (3.1 ish), 10% (2.8 ish)?
(I’ve done some googling, but can’t find any answers)
From https://zwiftinsider.com/crr/
A MTB would need to push 53 more watts than a road bike on tarmac (assuming a 75kg rider on a 7kg bike traveling at 40kmh).
You, are an absolute genius, guru, and all round top dude! Thanks so much for replying and putting all this information together…I’m sure i’m not the only one who is extremely grateful.
THANK YOU 🙂
I did the jungle loop this morning on the default mountain bike. I assume that was you on the leaderboard (who I beat 😛) . Average of 307w, time of 14:12. That’s over a minute slower than my PR which was set on a road bike at 201w average. From the times above it’s my PR which appears to be a bit off but IIRC I got that riding solo. I’ll try again on a road bike at over 300w, see what happens
That was our bot on the leaderboard, cooling off after a test up the Alpe. 😉
We’re still anxiously awaiting your comparison result, Jon.
Hi, If you enter a mtb race on zwift do you automatically get on a mtb or do you have to manually select it from your garage? Thanks!
Depends on how the event was set up, but generally, if they say it’s a MTB race, then they’ll force everyone to use a MTB.
It’s perhaps worth updating the results here with the test of the more recent Trek SuperCaliber: https://zwiftinsider.com/trek-supercaliber/ . I got the Trek for today’s Garmin-sponsored jungle “ride”, replacing my Zwift MTB. Small differences (5 seconds per lap, adjusting for my weight?) can be amplified (or compressed) in race situations, so I was glad I’d done the upgrade, barely winning the sprint (by a half-wheel) in my small chase group.
Added a link at the top to this post: https://zwiftinsider.com/updated-mtb/
Love riding in zwift on the mountainbike wish for Santa Cruz to enter the zwift world.