
Train for a 5K or parkrun PB with ChatGPT or Claude.
From your first 5K to a new parkrun PB, get personalised workouts, training plans and race pacing. Athmex keeps your recent times, goals and running schedule ready for ChatGPT or Claude. Build your free athlete profile to explore ways to use your latest parkrun result and recent runs from Strava or your watch.
You. Your fitness. Your activity data. Together in Claude and ChatGPT.
Can ChatGPT build a 5K training plan?
Yes, but “train me for a 5K” can describe very different problems. One runner is trying to complete the distance without stopping. Another runs a weekly parkrun and wants to stop fading in the last kilometre. Another is chasing a PB that has stood for three years.
ChatGPT or Claude needs your current race performances, available training time, recent running, pace or threshold information where you have it, and any constraints before its 5K advice becomes personal. Athmex gives the assistant that durable athlete context. Recent sessions can reach it separately, through your training platform.
- Distance
- 5 km
- In miles
- 3.1
- Saturday parkrun
- 5 km
Course, weather and start conditions all change what a 5K time means.
Bring your parkrun results and recent runs into your AI coaching
Your profile says who you are. Recent runs show what the last few weeks actually looked like.
Some training platforms can make recent activity available to ChatGPT or Claude through routes they support, so the assistant can check whether your easy runs have drifted fast or whether last Tuesday’s reps went to plan.
Athmex doesn’t import or store your activities. It keeps the athlete context behind them and helps you explore available ways to use your training information with ChatGPT or Claude. Build your free athlete profile to explore the options for your setup.
You. Your fitness. Your activity data. Together in Claude and ChatGPT.
Same request. Two very different 5Ks.
Both runners ask for the same thing. What they need from the next six weeks has little in common.
The same requestHelp me run a faster 5K.
Runner A
- Recent 5K
- 34:50, walking twice
- Running
- Finished Couch to 5K in spring
- Days
- 3 a week
- Constraint
- Shin soreness after hills
“Faster” here means running the whole way. The gain comes mainly from building easy, continuous running, with hills added slowly while the shins settle.
Runner B
- 5K PB
- 20:40, two years ago
- Recent parkrun
- 21:15
- Days
- 5 a week
- Goal
- Sub-20
A sharpening job with enough background already there. Sessions around threshold and 5K effort make sense, pitched from 21:15 rather than the old PB.
Identical words should still lead to different training. Choose the approach that matches your fitness and available training time.
What ChatGPT or Claude should know before setting your 5K training
A 5K plan is mostly a question of where you’re starting and how many days you have. These details set both.
- Current 5K
- Your PB and, more importantly, a recent 5K or parkrun time with its date.
- Goal
- Finish without walking, a round-number target or a specific PB.
- Running days
- How many days you can run and which ones allow a longer or harder session.
- Threshold pace
- If you know it. A recent race time works as a starting point if you don’t.
- Constraints
- Recurring injuries, work patterns, family commitments or hills you can’t avoid.
- Target race
- A specific parkrun, a club championship or a local 5K, with its date.
First 5K versus a faster 5K: two different goals
Completing a 5K is mainly about building enough easy, continuous running that 5 km feels manageable. A faster 5K keeps that base and adds work at and around race effort, which needs more careful recovery.
If you’ve only just finished a beginner programme, the Couch to 5K page covers that stage. This page picks up from there, whether the next target is running the whole way or a time that starts with a 1.
What determines useful 5K training advice
Three facts do most of the work: where your 5K is now, how much you’re running, and how many days you have. A plan for a runner on three days a week looks nothing like one for a runner on six, even at the same PB.
Constraints come next. A recurring calf strain, a job with early starts or a local area with no flat routes all change which sessions are realistic. An AI that doesn’t know them will plan around a runner who doesn’t exist.
Your current PB and recent fitness
A PB records your best day. Recent fitness is what the next block starts from. A 19:58 from four years ago and a 21:30 last month describe different runners, and pace targets built from the old number can leave every session feeling too hard.
Athmex stores PBs and current performances with the dates they were set, so ChatGPT or Claude can tell the difference. When a new time comes in, tell Claude, or ChatGPT on a supported plan, and ask it to update your profile.
Easy running, faster work and recovery in context
Good 5K training still rests on plenty of easy running. Faster sessions, such as reps at around 5K effort, short hill reps or a steady run near threshold, sit on top. The mix depends on the runner, and more hard sessions aren’t automatically better.
Recovery is where context matters most. A 45-year-old with a physical job and a 25-year-old with a desk job and free evenings recover differently. Your AI can only allow for that if it knows who you are.
Using parkrun as a goal or a benchmark
Standard Saturday parkruns are 5 km, which makes parkrun a natural first goal and a repeatable benchmark. Running the same course regularly gives you comparable times, as long as you allow for weather, mud and how busy the start was.
Courses differ, though. A flat tarmac loop and a hilly grass course can produce meaningfully different times for the same runner. If you add your home parkrun as a race or a note in Athmex, the assistant can read your times in context rather than comparing them with somebody else’s course.
From completion to PB
Many runners start parkrun hoping to get round and later decide they want a time. That’s a change of goal, and it’s worth recording. The training that got you round won’t necessarily take you under 25 or 20 minutes.
What a parkrun result tells you about current fitness
A recent parkrun is one of the most useful fitness markers a runner has, because it’s a real 5 km effort on a known course. It’s only a fair marker if it was a genuine effort, though. A time run as a social jog, pushing a buggy, pacing a friend or the morning after a hard session says something different from a time chased from the start line.
Read it with its conditions: the course, the weather, how busy the first kilometre was and how you felt in the last one. A runner who ran 24:10 with an even first and last kilometre is in a different place from one who ran 24:10 after going out at 4:30/km and fading. The first may have more to give at the same fitness. The second probably needs to learn to pace before chasing the time.
That’s also why a PB from two years ago makes a poor standalone target. It records your best day under the conditions of the time. Sessions pitched from it assume fitness you may not have now, which usually shows up as reps that can’t be held. Build from the recent time, and treat the old PB as the destination, not the starting pace.
Training example: from 24:10 towards a sub-23 parkrun
Consider a runner whose goal is a sub-23 parkrun in ten weeks. Their PB is 22:40 from two years ago. Their last three parkruns on the same flat course were 24:30, 24:20 and 24:10, and they run four times a week, about 25 km in total, with a session of reps on Tuesdays.
A reasonable conclusion: sub-23 is a realistic stretch, around 14 seconds per kilometre faster than now. Sessions should be pitched from 24:10, not from 22:40. A week with one session near 5K effort, one sustained run around threshold effort, a longer easy run and an easy run with strides would suit the goal and the four days.
Other context would change that. If the 24:10 came after a six-week break, fitness may come back faster than a plan built from scratch assumes. If the recorded runs show the last two Tuesday reps fading every week, the reps are pitched too fast. A recurring calf strain argues for fewer fast sessions and more easy running. And a target race on a hilly course makes 23:00 a different proposition.
The runs themselves come from the runner’s watch or training app. Athmex keeps the dated PB, the target, the four running days and the calf history, so ChatGPT or Claude reads those parkrun times against the right runner rather than against the old PB.
When pace zones or threshold information become useful
For a first 5K, perceived effort is enough. Once you want a time, pace information helps the assistant pitch sessions accurately. A recent 5K or 10K race time is a good starting point, and a threshold pace from a test or a coach adds precision.
Heart-rate zones can help too, especially for keeping easy runs easy. If you know your threshold heart rate or maximum heart rate, add it. If you don’t, leave it blank. Athmex keeps unknown values unknown rather than guessing.
Adapting when your training days change
A new job, a holiday or a child starting school can turn a five-day week into three. Ask ChatGPT or Claude to rework the plan and it should keep the sessions that matter most for your goal and drop what matters least.
Update your availability in Athmex when it changes and future conversations start from the new reality. If 5K is going well and you’re eyeing a longer race, the 10K training page covers what changes when the distance doubles.
Where AI 5K coaching has limits
An AI can misread a note or overestimate what you can handle, especially when you ask for a big jump in speed. A 5K plan that adds hard sessions too quickly is a common route to a calf or Achilles problem.
If something hurts in a way that changes how you run, get it looked at by a physio or doctor rather than asking the assistant for a harder week. It can’t examine you, and it shouldn’t be treated as if it could.
Questions runners ask
Can ChatGPT make a 5K training plan?
Yes. ChatGPT or Claude can plan a 5K block and explain each session. The plan is more suitable when it knows your recent 5K time, running days, goal and constraints, which Athmex keeps ready.
Can AI help me improve my parkrun time?
It can help plan and adapt the training. Give it a recent parkrun time, which course you run, how many days you train and your target, and it can pitch sessions from where you are now.
What should ChatGPT know before setting my 5K training?
A recent 5K time and its date, your goal, how many days you can run, threshold pace or heart rate if you know them, and any injury or constraint you train around.
Is my parkrun time a good measure of my 5K fitness?
A recent parkrun run as a genuine effort is one of the best markers you have. Read it with its course, weather and pacing: a fast start and a fade, a crowded first kilometre or a muddy course all change what the time means.
Give ChatGPT or Claude the 5K context it’s normally missing.
Build your athlete profile once. Connect ChatGPT or Claude. Keep your recent times, running days, goal and constraints ready every time you ask about the next session.
You. Your fitness. Your activity data. Together in Claude and ChatGPT.