Why don't I see data for a model?

You expected to see responses from a model, Gemini, Perplexity, Claude, and it shows nothing. Four causes explain nearly every case, and the first is by far the most common.

​

1. The model isn’t selected on that monitor

Models are chosen per monitor, not per project. Every prompt in a monitor runs against exactly the models that monitor has selected, so a model you never added produces no data, however prominently it appears in the dashboard’s filter list. It’s easy to assume model tracking is global; it isn’t, by design, because each monitor represents one collection context.

​

How to check

Open the monitor, go to its Settings, and review the Models grid; add the missing model if you want it tracked. Repeat per monitor if you have several: selecting Gemini on your US monitor doesn’t add it to your German one. See choosing your models and what is a monitor.

​

2. Your plan doesn’t include the model

The entry-level Explore plan tracks ChatGPT only. Every paid plan unlocks all active models. On Explore, the create-monitor form shows other models with an Upgrade badge and won’t let you check them, so “no data for Gemini” is the plan working as designed.

​

How to check

On the create-monitor form, gated models carry an Upgrade badge. In an existing monitor’s Settings they look selectable, but saving a disallowed model fails with a plan error; that save error is plan gating. See plans and billing.

​

3. The model was added recently, so there’s no history

Data collection starts when a model is added and is never backfilled: responses only exist from runs that actually included the model. When you add a model to an active monitor, Promptwatch fast-tracks its first responses rather than waiting for the next scheduled cycle, so the first data typically lands the same day. But a chart scoped to last month will still show nothing for it, correctly.

​

How to check

Narrow the date range to the last few days. Data present recently but absent before the model was added is the expected shape, not a gap. The same applies in reverse to retired models: when a provider replaces a model version, the old version stops collecting but its historical responses stay in your data.

​

4. Today’s run hasn’t reached it yet

Prompts run on the monitor’s schedule, daily by default, and responses arrive in batches through the day rather than all at once. Each response is then analyzed before it counts in mention and visibility charts. Looking at a “today” view mid-morning, it’s normal for some models to have reported while others haven’t.

​

How to check

Wait for the day’s cycle to complete, or widen the range to include yesterday. If a model shows nothing for several complete days while others in the same monitor report normally, that’s no longer cadence, and worth reporting to support. See how fresh is my data for the full collection timeline.

​

When empty is expected

A model that is unselected, plan-gated, brand new to the monitor, or mid-cycle is supposed to look empty. Ruled all four out across several days? Check that the monitor is Active, then check your organization’s response usage under Settings → Billing. A paused monitor stops that monitor only; an exhausted monthly response quota stops new collection for every monitor until the period resets or the limit is raised. See managing your response limit.