---
title: AI Gateway Model Allowlist
product: vercel
url: /docs/ai-gateway/security-and-compliance/model-allowlist
canonical_url: "https://vercel.com/docs/ai-gateway/security-and-compliance/model-allowlist"
last_updated: 2026-09-08
type: conceptual
prerequisites:
  - /docs/ai-gateway/security-and-compliance
  - /docs/ai-gateway
related:
  - /docs/ai-gateway/security-and-compliance/provider-allowlist
  - /docs/ai-gateway/models-and-providers/virtual-models
  - /docs/ai-gateway/security-and-compliance/private-inference
  - /docs/ai-gateway/authentication-and-byok/byok
  - /docs/cli/ai-gateway
summary: Restrict which AI models your team can use through AI Gateway. Available on Pro and Enterprise.
install_vercel_plugin: npx plugins add vercel/vercel-plugin
search_with_vercel_cli: npx vercel@latest docs search "<query>"
---

# AI Gateway Model Allowlist

The model allowlist lets team owners restrict which models can be used through AI Gateway. A request returns a `403` when the requested model is not on the allowlist. The feature is opt-in and available on Pro and Enterprise plans.


<!-- docsgraph:related -->
## Related pages

> **For AI agents:** Follow these links to understand how this page connects to the rest of the Vercel ecosystem. For the full cross-link map (inbound, outbound, prerequisites, and semantic neighbors), see the .graph.md link below.

- [How to control AI provider data handling with AI Gateway](https://vercel.com/kb/guide/control-ai-provider-data) — Prevent AI providers from training on your prompts, enforce zero data retention, and restrict providers and models with
- [Team-wide provider allowlist on AI Gateway](https://vercel.com/changelog/team-wide-provider-allowlist-on-ai-gateway)
- [AI Gateway Models and Providers](https://vercel.com/docs/ai-gateway/models-and-providers) — Choose AI Gateway models and providers. Configure routing, fallbacks, timeouts, prompt caching, reasoning, and web searc
- [AI Gateway FAQ](https://vercel.com/docs/ai-gateway/faq) — Answers to common questions about AI Gateway, including request errors, pricing and markup, caching, SDK and API compati
- [AI Gateway Pricing](https://vercel.com/docs/ai-gateway/pricing) — Understand AI Gateway token pricing, free and paid credits, BYOK costs, add-on charges, and payment fees. Manage credit
- [AI Gateway Disallow Prompt Training](https://vercel.com/docs/ai-gateway/security-and-compliance/disallow-prompt-training) — Learn how to prevent AI providers from using your prompts and responses for model training through AI Gateway.
- [List virtual model configs](https://vercel.com/docs/rest-api/ai-gateway/list-virtual-model-configs) — GET /ai-gateway/virtual-model-configs/list — List virtual model configs. For a team, returns all of its VMCs, or with \\`

Full cross-link map for this page: [/docs/ai-gateway/security-and-compliance/model-allowlist.graph.md](/docs/ai-gateway/security-and-compliance/model-allowlist.graph.md)
<!-- /docsgraph:related -->

A common reason to enable this is governance: your team has approved a specific set of models, and you want to guarantee that AI Gateway never serves one you haven't reviewed.

## Pricing

| Option                    | Cost                                | Availability       |
| ------------------------- | ----------------------------------- | ------------------ |
| Team-wide model allowlist | $0.10 per 1,000 billed requests     | Pro and Enterprise |

The surcharge applies to each request AI Gateway bills. That includes a request billed without completing normally, such as a response the provider blocks on safety grounds and still charges for. Requests blocked by the allowlist (`403`) and other failures that AI Gateway doesn't bill are not charged.

If your team also has the [provider allowlist](/docs/ai-gateway/security-and-compliance/provider-allowlist) enabled, a request is charged a single restriction surcharge, not one per allowlist. You are never billed twice for having both active.

## Enabling the allowlist

Only team owners can change this setting.

1. Open the [AI Gateway **Settings** tab](https://vercel.com/d?to=%2F%5Bteam%5D%2F%7E%2Fai-gateway%2Fsettings\&title=AI+Gateway+Settings).
2. Toggle on **Model Allowlist**.

When you toggle on, every model currently available on AI Gateway is added to the allowlist by default. This guarantees that turning the feature on never breaks existing traffic. You then remove models you want to block as an explicit action.

## Disabling specific models

With the allowlist enabled, switch off any models your team should not use. Models are grouped by their creator.

When a request targets a disabled model, AI Gateway returns:

```json
{
  "error": "Your team has restricted access to this model. Team owners can edit AI Gateway Settings (https://vercel.com/d?to=%2F%5Bteam%5D%2F%7E%2Fai-gateway%2Fsettings&title=AI+Gateway+Settings) to enable this model in Model Allowlist. Model: example/model-name",
  "type": "no_providers_available",
  "statusCode": 403
}
```

## New models

When AI Gateway adds a new model after you enable the allowlist, the new model is disabled by default and shows a **New** badge in the settings list. An info note at the top of the list tells you how many new models are waiting for a decision.

To clear the badges, decide which new models to enable and click **Save**. A save with no changes also counts as acknowledgment and clears the badges. Disabled new models stay disabled until you enable them.

## Virtual models

A request through a [virtual model](/docs/ai-gateway/models-and-providers/virtual-models) is evaluated against the underlying model it resolves to. If that model is not on the allowlist, the request is blocked, even though the virtual model itself is not listed in the settings.

## Private models

The allowlist applies to models in the AI Gateway model library. Your team's [private models](/docs/ai-gateway/security-and-compliance/private-inference) are endpoints you registered yourself, so they remain available and are not affected by the allowlist.

## Provider allowlist

The model allowlist and the [provider allowlist](/docs/ai-gateway/security-and-compliance/provider-allowlist) are independent checks, and a request must pass both:

- The model allowlist controls which models your team can use.
- The provider allowlist controls which providers can serve them.

A model on your allowlist can still be blocked if every provider that serves it is disabled on the provider allowlist. Keep at least one of its providers enabled for each model you want to use.

If both allowlists are enabled, you are charged a single restriction surcharge per billed request, as described in [Pricing](#pricing).

## BYOK

The allowlist applies to [BYOK](/docs/ai-gateway/authentication-and-byok/byok) requests too. Disabling a model means your team does not want any traffic to that model through AI Gateway, even with your own credentials. The [surcharge](#pricing) applies to BYOK requests too, charged to your AI Gateway Credits.

## API keys that bypass the allowlist

Team owners can create an API key that's exempt from all of the team's restrictions: this allowlist, the provider allowlist, ZDR-only models, and the HIPAA and prompt-training filters. Create it with the Vercel CLI, and give it an expiration so the exemption doesn't outlive the work that needed it:

```bash filename="terminal"
vercel ai-gateway api-keys create --name escape-hatch --bypass-all-settings --expiration 7d
```

See [`api-keys create`](/docs/cli/ai-gateway#create) for every option.

## Plan and role requirements

| Requirement | Detail                            |
| ----------- | --------------------------------- |
| Plan        | Pro or Enterprise (not Pro trial) |
| Role        | Owner                             |

Team members on non-owner roles can view the current configuration but cannot modify it.

> **💡 Note:** Allowlist changes can take a couple of minutes to propagate. In-flight
> requests finish under the previous settings. Once the change has propagated,
> new requests respect the updated allowlist.

## Search the docs from your terminal

Search the Vercel docs and read pages as Markdown with the Vercel CLI. No login required.

```bash
# Find docs pages by keyword
npx vercel@latest docs search "environment variables"
npx vercel@latest docs search "function timeout"

# Print a docs page as Markdown, by path or URL
npx vercel@latest docs inspect <page>
```

---

[View full sitemap](/docs/sitemap)


If you encounter incorrect information, can’t find what you need, need a capability Vercel doesn’t currently offer, or have any other feedback to report, please do it by following the instructions at https://vercel.com/feedback.md.
