Swap every AI in your hive to a coordinated model strategy with one click — Cheap, Balanced, Thinker, Fast, or a configuration you've saved yourself. This page explains how the built-in profiles pick models, when to save your own, and what the override badge is telling you.
The Hive Profile dropdown sits at the top of the Worker Bees screen (Internet mode). Picking a profile from it tells every AI in your hive to switch to the model that fits that profile's intent — without you having to click into each AI row and pick models one at a time.
Five profiles ship built-in (🛠 Custom plus four tier-based picks), and you can save as many of your own as you want.
| Profile | What it picks | Use when |
|---|---|---|
| 🛠 Custom | Whatever's already saved on each row | You've hand-picked specific models per AI and want them left alone |
| 💰 Cheap | Cheapest model each provider considers reliable for production use | Long sessions, big documents, draft work where token cost matters more than polish |
| ⚖️ Balanced | Best capability per dollar — the provider's "default" answer | Your everyday hive. Good middle ground for most documents |
| 🧠 Thinker | Deepest reasoning model each provider offers | Complex analytical documents — RFPs, legal review, technical guides where structured thought matters |
| ⚡ Fast | Lowest-latency model each provider offers | Quick iteration, large hives where round-time stacks up, drafts where speed beats nuance |
When you pick a tier profile (Cheap, Balanced, Thinker, or Fast), WaxFrame doesn't guess which model is which — it asks each provider's own API to classify its models. Each AI receives a numbered list of every model it offers and is asked to pick one for each of the four tier slots, plus a one-sentence reason for each pick.
The classification gets cached per provider in your browser. The next time you pick a tier profile, the answer is already there and the dropdown applies instantly with no API calls.
If you pick a tier profile and one of your providers hasn't been classified yet, you'll see a toast like "💰 Cheap — classifying 2 providers in the background, picks will fill in shortly…"
WaxFrame fires the classification calls in the background. As each provider's classification lands, the profile re-applies silently — the new picks get swept into the hive without a second click from you.
If you flip to a different profile while classification is still running, the re-apply is skipped — your new choice wins.
When a Hive Profile is active and you manually swap one AI's model — say you applied the 💰 Cheap profile but then changed Claude to the Opus model for a particular session — that AI's row gets an ✏️ override badge next to its model dropdown.
The badge tells you, at a glance, which AIs are now diverging from the profile's intent. Hover the badge to see what the profile would have picked instead:
"Claude is using claude-opus-4-5 — the active 💰 Cheap profile would pick claude-haiku-4 instead. Switch the model from the row above to bring it back into the profile."
Switching the model back to what the profile expects clears the badge. Switching the profile to Custom also clears it (Custom has no expected pick to diverge from).
Built-in tier profiles are great for general intent ("always pick the cheap one"). But sometimes you want to preserve a SPECIFIC configuration — a hive you've hand-tuned for a particular kind of work — so you can pull it back up with one click later.
Click 💾 Save current as profile in the Hive Profile bar. A modal asks for a name (40 chars) and an optional single-emoji icon. The current per-AI model picks are captured as a snapshot under your chosen name.
Your saved profile shows up in the Hive Profile dropdown under a "Your saved profiles" group, separated from the built-in tiers. Picking it sweeps every AI back to the model it had at save time.
Use a built-in tier when you want a category that tracks new models as they appear. "Cheap" in six months will pick whatever the cheapest reliable model is then, not the model that was cheap today.
Use a saved profile when you want exactly these models, in exactly this combination, forever. If a provider releases a new cheaper model after you save, your saved profile keeps the old pick — that's the point. Snapshot, not category.
Saved profiles capture an AI's model snapshot, not the AI itself. If you delete a Custom AI from your hive and then apply a saved profile that referenced it, that AI is silently skipped — you'll see a count in the toast ("Applied to 8 of 10 AIs, 2 missing or unkeyed").
If an AI's key gets removed but the AI is still in the hive, same outcome: skipped, counted in the missing-or-unkeyed total.
Saved profiles only capture AIs that had a key and a non-empty model at save time. AIs you add later won't be touched by the profile — they keep whatever model you've picked for them. To bring the new AIs under the profile's coverage, save a fresh profile that includes them.
If a saved profile points an AI at a model that the provider has since retired, the next time you apply the profile the row will land on a model that no longer exists. You'll usually see this as a ⚠ deprecation flag on the AI row. Switch the row to a current model and re-save the profile to keep it useful.
Enter your license key to continue.
Don't have a key? Buy WaxFrame Pro(opens in a new tab)
Your WaxFrame Pro license is active.
Hive Profiles