Ask Maxiom’s chat coach, Max, when you’ll peak for Ironman. It hands you a confident number. Ask how it got that number, and the conversation stops.
AthleteOS wins on transparency. Every fitness score, fatigue score, and form score comes from one documented formula. You can recompute it yourself. Maxiom wins on conversation. Its chat interface is genuinely well built. A 2026 hands-on test of AI triathlon apps said it delivers key numbers without burying you in dashboards. Both things are true. This piece is about which one matters more.
Quick map of what’s below:
- Want a coach that texts like a person? See “What Maxiom’s Max Actually Does.”
- Want the math behind your fitness number? See “The Math Behind the Numbers.”
- Worried about ramping load too fast? See the ACWR section.
- Wondering if AI can write a safe plan alone? See the AI-plan section.
- Deciding on price? Jump to the pricing table.
AthleteOS vs Maxiom at a Glance
| Feature | AthleteOS | Maxiom (Max) | Edge |
|---|---|---|---|
| Core interaction | Dashboard, calendar, AI coach | Chat, voice | Different styles |
| Load metric shown | Fitness, fatigue, form scores | ”Recovery score,” race predictions | AthleteOS |
| Formula published | Yes | No (July 2026) | AthleteOS |
| TSS source | Formula-based | Undisclosed | AthleteOS |
| Load chart recomputable | Yes | No | AthleteOS |
| Ramp-rate ceiling | Tracked, visible | Undisclosed method | AthleteOS |
| Recovery-score method | Disclosed | Undisclosed | AthleteOS |
| Primary UI | Dashboard-first | Chat/voice-first | Maxiom |
| Wearable / sleep sync | Yes | Yes, + Apple Watch app | Tie |
| Athlete endorsements | None | Blummenfelt, Iden, Philipp (Maxiom’s claim) | Maxiom (marketing) |
| Clinical validation | Not claimed | ”Entering trials” (2025 release); no results | Neither proven |
| Pricing (verified) | Free + paid tiers | $20-$100/mo, iOS, July 2026 | See pricing |
The Math Behind the Numbers: CTL, ATL, and TSB, Explained
AthleteOS’s load model isn’t unique to us. It’s the Banister impulse-response model. Coggan and Allen later called it the Performance Management Chart.
Fitness score (CTL) = 42-day exponentially weighted average of daily training stress
Fatigue score (ATL) = 7-day exponentially weighted average of daily training stress
Form score (TSB) = Fitness score - Fatigue score
In plain English: fitness score is a slow six-week average of training. Fatigue score is a fast one-week average. Form score is the gap between them. High fitness, low fatigue means positive form. You’re fresh. Load hard for a week and fatigue spikes faster than fitness. Form turns negative. That number says one thing: you’re tired.
Maxiom’s own blog post cites a CTL target of 190, up from 100. That’s one company post, not independent proof. Still, it shows Max is built on load concepts AthleteOS shows openly. What Maxiom doesn’t do, as of July 2026, is publish the formula or explain how CTL feeds its recovery score.
That gap matters. TrainingPeaks’ guidance puts a safe weekly ramp at 5 to 8 points. Running stays more conservative, 3 to 5, given injury risk. Going from 100 to 190 in two months blows past that. Whether Max respects any limit isn’t something you can check.
The chart isn’t proof Maxiom does anything wrong. It’s proof nobody outside the company can check. AthleteOS shows the same numbers on the same chart. You could rebuild the formula in a spreadsheet.
What Maxiom’s Max Actually Does
Give credit where it’s due. MAXIOM Labs launched Max in May 2025, calling it “the world’s first human-centered AI coach for health & performance.” It combines wearable data, biometrics, and chat coaching. Real features: voice check-ins, a recovery score, race predictions, photo meal logging, and an Apple Watch app.
A 2026 hands-on test of eight AI triathlon apps included Maxiom. Testers called it a strong training hub that surfaces key numbers without clutter. Workouts adjusted to recovery and check-ins, though not always perfectly. That review sits behind a paywall, so treat this as directional.
Pricing, from the iOS App Store in July 2026: free download, then Core Monthly at $20, Core Plus at $30, Premium at $100. Annual options run up to $1,000. Pricing shifts often, so check the current listing.
Is the Acute:Chronic Workload Ratio Real Science?
Maxiom’s marketing, like a lot of coaching content, leans on load-ratio language as if it’s settled physics. It isn’t.
Gabbett’s 2016 paper found a ratio of 0.8 to 1.3, recent week versus six-week average, sits in the lowest injury-risk zone. Ratios above 1.5 flag higher risk. In short: don’t spike your load faster than your body adapts.
That’s the popular version. The critical version is messier. A 2020 review of 27 studies found 0.8-1.3 trended toward lower injury risk, but quality hit just 59.7%. A sharper critique, from Impellizzeri and colleagues in the International Journal of Sports Physiology and Performance, argues the ratio’s statistics are shaky and its time windows are arbitrary. A related paper found the standard formula inflates correlations on its own, from the math alone.
None of that makes load tracking useless. It means one ratio isn’t a hard verdict on injury risk. AthleteOS shows ACWR next to fitness, fatigue, and form, as one signal among several. The same logic applies when rebuilding load after time off. In fact, the classic 10% weekly mileage rule turns out to be the wrong tool for it.
What Happens When You Ask an AI to Write a Training Plan
This research should make anyone pause before trusting a chatbot with full training authority.
In one study, 10 coaches rated ChatGPT running plans on 22 quality criteria. Minimal-prompt plans scored a median of 2 out of 5. Plans built from detailed athlete data scored a median of 4 out of 5. Even that best version wasn’t rated optimal.
A second study asked 10 coaches to rate AI strength plans across 27 criteria. Only 5 criteria scored 4.5 or higher. Worse: agreement between coaches was essentially zero, a Fleiss’ kappa near 0. They couldn’t agree on how good the plans were.
That’s not a knock on any one AI. It shows plan quality is hard to audit without a transparent model. That’s why AthleteOS splits the job in two. The AI coach designs the workout. A separate, deterministic calculator scores it into training stress. It never guesses that number. The AI proposes. The formula measures.
Black Box vs Glass Box: Why Recomputing the Number Matters
A 2024 review of explainable AI in clinical care found transparency raises trust, but only when the explanation is simple. Confusing explanations backfire. That’s a caution against over-explaining, and it applies here too. The formula behind your fitness score isn’t a black box, and it isn’t jargon either. It’s arithmetic: six-week average minus one-week average. You can check it with a calculator.
Maxiom’s recovery score and race predictions might be excellent. They might even beat a plain CTL/ATL/TSB model sometimes. Nobody outside the company can say, because the model isn’t public. That’s the entire difference this comparison rests on: whether you can check its work.
Pricing: What Each Actually Costs
| Tier | Maxiom (iOS, July 2026) | AthleteOS |
|---|---|---|
| Free | Download only, limited features | Free tier available |
| Entry paid | Core Monthly: $20/mo | See current tiers |
| Mid paid | Core Plus Monthly: $30/mo | See current tiers |
| Top paid | Premium Monthly: $100/mo | See current tiers |
| Annual entry | Core Annual: $192.99/yr | See current tiers |
| Annual top | Premium Yearly: $1,000/yr | See current tiers |
AthleteOS pricing changes over time. Check current AthleteOS pricing rather than trusting a number frozen in an article. The same goes for Maxiom: App Store prices get revised often, and these figures are accurate only as of writing.
Use AthleteOS If… Use Maxiom If…
Use AthleteOS if:
- You want to recompute the formula behind your numbers.
- You train multiple sports and want one load model.
- You’ve been burned by a “just trust the app” claim before.
Use Maxiom if:
- A chat-first interface fits how you like to train.
- You value photo meal logging and community features.
- You’re fine trading transparency for a conversational feel.
What We Can’t Do That Maxiom Can
Fair is fair. Maxiom’s chat interface is a genuinely different product, not just a dashboard skin. A daily nudge that reads like a text from a person is a real usability win. So is photo-logging a meal instead of searching a database. AthleteOS doesn’t match those features yet. If chat-first interaction keeps you engaged, that’s a fair reason to pick Max.
A Quick Example
Take a fictional athlete: Dev, 34, training for his second Ironman on 9 hours a week. He tried a chat-based AI coach for six weeks and liked the check-ins. Then his “recovery score” dropped from 82 to 61 in one day, with no explanation. Was it sleep? A hard bike session two days back? He couldn’t tell. A formula-based load model breaks that same drop into fatigue score, sleep, and HRV trend, each traceable to a session.
Where AthleteOS Fits
If you’re comparing AI platforms, you likely care about the number behind your next workout. Not just receiving it. That’s what AthleteOS’s Performance Management Chart is built for. Read more on what CTL, ATL, and TSB actually mean. See how AI coaching apps compare in 2026. And why Zone 2 volume still matters more than any AI number on top. You can also try AthleteOS and see the formula yourself.