Run notes · 13 July 2026

Hevy and Strong gained on the leader. Fitbod's gain didn't hold up.

I found fitness had no favorite in the first snapshot. The second one shows the fragmentation isn't static. Two apps are gaining on the leader for real. A third only looked like it was, until I checked which engines actually moved.

Update, 3 August 2026: this didn't hold. The August snapshot reversed it. On ChatGPT, Hevy fell from 17 percent back to 6 and Strong from 17 back to 6, both at or below where they started in June, while Nike Training Club rose from 17 to 26 and extended its lead. The gap didn't close. What I read as two challengers moving together was a one month swing that came back. I've left the piece as written, because the reasoning it applies is the reasoning that failed, and that turned out to be the useful part. See how much of a month's movement is real for what three months of data say about calling a move this early.

Run note Written by Claude from that month's data, and read by me before it went up. The numbers come from the index, not from the model. My own reading of the data is in Analysis.

Fitness has no favorite: the leader, Nike Training Club, is named in about a third of grounded answers, with the rest of the field close behind. That was true again in July. What's new is that "close behind" now has a direction. Hevy and Strong, both lifting-tracker apps built for a narrower job than Nike Training Club's general workout library, gained on the leader across the engines that carry the most samples, not just the smaller capture.

Hevy's ChatGPT rate went from 9% to 17% (3 of 35 answers to 6), and Claude from 17% to 23%. Strong's ChatGPT rate went from 6% to 17%, Claude from 29% to 34%. Both engines sample 35 questions, so a move like that is a real shift in how often the model reaches for the app, not a rounding artifact. Nike Training Club's overall score barely moved, down one point, but that flatness hides its own crosscurrent: Claude named it much more this month, 20% to 34%, while ChatGPT and Google's AI Overview both named it a little less. The leader isn't sliding. Its gains and losses are canceling out while two challengers gain on multiple fronts at once.

The category didn't get a new leader this month. It got two apps closing the gap on more than one engine at a time, which is a different, more durable kind of movement than a single column swinging.

Why Fitbod isn't on this list

Fitbod's score also went up, from 22 to 28, the same size move as Strong's. I checked it the same way and it doesn't hold up. ChatGPT actually went down three points for Fitbod this month, Claude was flat, Gemini down three, Mistral up three. On the four engines that each sample 35 questions, Fitbod is a wash. The entire six-point gain traces to Google's AI Overview, which names Fitbod in one of its six monthly captures in June and three of six in July. That's a real change in what that specific capture returned, but it's one extra mention out of six samples, and I'm not willing to call that a trend the way I will for Hevy and Strong's moves on 35-sample engines. It's possible Fitbod is gaining too and the API engines haven't caught up yet. It's also possible this is exactly the kind of swing that reverses next month. I don't have enough to say which.

One month isn't a trend either

Two data points make a line, not a pattern. If Hevy and Strong keep gaining in August, that's a real story about the strength-training corner of fitness pulling ahead of the generalist app. If they give it back, June-to-July was noise that happened to be broad instead of narrow, and broad isn't the same guarantee as repeated. I wrote up the general version of this check, and where it did and didn't hold across every category this month, in how much of a month's movement is real.

What this is

Thirty-five questions put to four AI assistants, plus Google's AI Overviews captured from the web, run in June and again in July on the same frozen prompts. US English. See the fitness index for the current board.