“Col de la Croix de Fer” Details (Zwift Climb Portal)
Key Details
- Category: HC
- Length: 27.1 km (16.8 miles)
- Elevation Gain: 1558 m (5,112‘)
- Average Gradient: 4.7%
- Country: France
- Watopia Climb Segment
- Watopia Descent Segment
- France Climb Segment
- France Descent Segment
- 5 W/kg: 68 minutes
- 4 W/kg: 83 minutes
- 3 W/kg: 103 minutes
- 2 W/kg: 148 minutes
Strava Segments
Time Estimates
Located in France’s Rhône-Alpes, climbing Col de la Croix de Fer (Pass of the Iron Cross) is a truly unique experience with its varied views and pitches. Don’t be fooled by its average gradient: the profile is irregular, with flowing sections, steep ramps over 12%, and even some downhill bits. You’ll have to constantly adapt and be strong both physically and mentally to complete it.
In 2026, the Tour de France will tackle this climb for the 22nd time. It’s the first climb in stage 20, the Tour’s epic Queen Stage, which also includes Col du Télégraphe and Col du Galibier (Valloire).
The TdF website says this climb, “perfectly embodies the toughness and heritage of the Tour de France.” While the pass can be tackled from multiple directions, Zwift’s version begins to the south in Allemond, since that’s what the pros are riding in 2026.
A bit of Zwift trivia: When it was added to the game in May of 2026, this climb became the longest in the Portal library, in terms of completion time estimates. While Mt. Hamilton is a few kilometers longer, Col de la Croix de Fer includes an additional 300 meters of elevation!
Here’s the verified IRL Strava segment:
This ALMOST makes up for Zwift never releasing the Stelvio. Almost. Nothing like a new longest climb to get me excited.
I rode the Col de la Croix de Fer this afternoon and it was a worthy challenge for sure. The long ramps of 6,7-8% seemed to go on forever in the middle part of the climb. Thank you Zwift for releasing this one and hopefully the Stelvio Pass someday as well.
I wonder if with AI, Zwift can generate generic roads and a more natural looking environment which, although not true to life, would be far superior to the Tron-like ‘world’ the portal rides exist in currently (which aren’t bad btw, just some trees and such would be nice).
I hope that they will use the Rouvy purchase to eventually rework the climb portal in Zwift. As it is right now, I cannot look at this for more than 20 minutes.
Is there a way to estimate the time at 50%?
I tried at 50% earlier this week and could only complete 60% around the hour mark (including lead in), had to log off for work. My estimated finish time was 100ish minutes or so based on 1.6W/kg.
great thanks. so wild guess is 80ish minutes at 2 w/kg
Once you start there is an ETA displayed in the lower right.
I am ~150 FTP, doing the climb at 50%, 90 watt average: ~2 hr 30 minutes up, ~1 hr down. The upgrades on the downhill are fun.
A couple of folks were messaging after the first hour about much further they had to climb.
“It was just a 2 hour hill climb” Gilligan’s Island “3 hour tour” reference.
I also had a 1 hr break after the first hour – I had to go out on a fire/rescue call.
You can add a little bit of spice when also going back after the summit.This adds about 300m of height with some very steep climbs. That’s different to many other climbs where you can ‘let your bike go downhill on its own’.
Looks like there is another 260m of climbing on the descent alone, if you want to complete that too!!
Did this climb IRL years ago. I’d forgotten how hard it was. And the descent isn’t easy. Keep something in reserve if you want to get down.
I tapped out on the way down after the second big climb. I just didn’t have it in me. I knew on the way up that those parts would suck on the way down.
Yeah, I saw those on the way up and decided early that I wasn’t hanging out for the “descent.”
I couldn’t imagine riding it IRL at altitude… I generally enjoy breathing. Lol
Congrats 🎉
Good to see that there are some new real climbs. Some calculation problems however: The title says a) 27,4 km at 4,7%. In the key details, they mention b) 27,1 km at 4,7% and the elevation gain is 1558m. If you use a), the average gradient would be 5,69%. If you use b), the average gradient would be 5,75%. A bit of consistency and good calculation would be appreciated, this is not rocket science.
I appreciate your comment and its attempt to help, until that final sentence. So close! 😅
Elevation gain is the total amount of climbing you do. That’s a different figure from elevation difference, which is the difference between your starting elevation and your ending elevation.
Average gradient isn’t computed using elevation gain. It’s computed using elevation difference. So the numbers above are correct.
I sincerely apologize Eric for my last sentence. There’s a substantial difference between my calculation and the elevation gain in this case because of a few downhill sections in this climb. However, the title indicates a length of 27.4 km and the key details 27.1 km. By the way, thanks 100 times for adding 5 new climbs to the portal (Alto de Patios, Col de la Croix de Fer, Col de Sarenne, Col du Galibier (Valloire) and Col du Télégraphe. I’ve done all 4 versions of all the other climbs except Passo dello Stelvio and Sa Calobra which I’ve never seen open. Are there on the master list or not? My apologies again. Éric
No worries, Eric. I appreciate the apology.
The 27.4km you’re seeing… I assume you’re talking about what the Veloviewer (and Strava) show for distance? It’s pretty common for Strava’s distance amount to differ from what Zwift shows in game, so what I do is use Zwift’s figure, since that’s what you’ll actually see when riding it.
Stelvio and Sa Calobra both exist in Zwift, but have never been scheduled, so I don’t include them in my master list. My guess is there’s a bug somewhere in those climbs, and we won’t see them unless/until Zwift fixes it.
Thanks for your reply Eric. To answer your question, the title indicates a length of 27.4 km, the Key Details 27.1 km and the verified IRL Strava segment 27.6 km, all on this page. I can understand that the Strava segment is different from Zwift but I would expect the first two numbers to be the same. Regarding the current master list, it appears that both Stelvio and Sa Calobra are still there (total of 54 climbs now): https://whatsonzwift.com/climb-portals. You won’t be surprised that my background is in mathematics and statistics!
Did the 100 % and the 125 % b2b.
I really enjoyed the climbs, long sections, where you dont have to change gear a lot. Its a great pacing exercise and more mental then for the legs. 1558 m didnt feel like it to be honest.
The downhills however…holy moly. At 100 % it was like 270 m of elevation gain, at 125 % something like 320 m. They acutally have the steepest sections, with 25 % in the 100 %.
So all in all, loved the uphill, hated my self for going downhill. After Mont Ventoux my favorite climb. Thnx for adding Zwift.
If you prefer not to shift may I recommend setting the trainer difficulty to 0%?
Checked this climb off off last night, you are not joking about required physical and mental toughness. The accent and decent (if that’s what you want to call it) took 3 hours.
Just completed the uphill at 100% in 1:51.21. Definitely a mental challenge on top of the physical. Similar to doing a real century, not quite the same as doing the Four Horsemen. The return is cray-cray… there are multiple climbs with double digit gradients, including one that flips up to 22%. Quit at the top, save something for the return, or take a long break at the top before you start.
According to Strava, I’m the second man in my weight class to even bother attempting this, and I was a bit faster than the other 102-113 kg madman.
Rode this at the beginning of the month as preparation for the IRL L’etape du Tour (along with the other 3 climbs (Telegraph, Galibier and Sarenne). Whilst I’m not a fan of the portal, I can’t credit it enough for the mental preparation ahead of the event. A good representation of the the real climb although would be far more enjoyable with some scenery as others have mentioned (not a fan of the portal as is).