Accessibility Adjustments

Use these optional tools to adjust reading and display preferences. These tools cannot resolve every accessibility barrier. Please contact the website owner if you need assistance.

  • Text adjustments
  • Content scaling 100%
  • Font size 100%
  • Line height 100%
  • Letter spacing 100%
  • Colour adjustments
  • Orientation adjustments

Vercel Token Pass raises monthly credit to $150, with fallback costs

Listen to this article

Vercel raised Token Pass to $150 in monthly model credit for $30 per team on November 13, 2026. The larger allowance replaces $80, but it does not make every AI Gateway model cheaper or put a hard ceiling on every connected agent’s spending. The credential used by the application changes what happens when the allowance runs out.

The November 13 announcement describes an early-access add-on for Pro teams, with limited places and a waitlist when capacity fills. The $30 charge is additional to the underlying Pro subscription. This is a credit-package change, not a new model launch.

Which models the credit covers

The current Token Pass documentation lists Kimi K3, Kimi K3.1 Pro, GLM 5.3, GLM 5.3 Flash and DeepSeek V4.1 Flash. Coverage requires the featured Token Pass provider, currently served by Wafer. A model appearing elsewhere in AI Gateway does not automatically make that route eligible.

Credit is consumed at each model’s listed token rates, rather than buying a fixed number of prompts or completed tasks. The US dollar rates list Kimi K3.1 Pro at $2 input and $10 output per million tokens, compared with $0.15 input and $0.60 output for DeepSeek V4.1 Flash. Cached-input rates are separate.

For perspective, one million uncached input tokens and one million output tokens would draw $12 of credit with the listed Kimi model, or $0.75 with DeepSeek. That is arithmetic from the published rates, not a benchmark or a prediction of equal work. Models can produce different amounts of reasoning, require different retries and finish with different results.

A scoped key stops where a standard key can keep spending

Vercel’s November 6 routing change lets eligible usage draw Token Pass credit automatically through standard API keys, OIDC authentication or fx login. A separate Token Pass key is optional, not a prerequisite for receiving the allowance.

The distinction matters after exhaustion. A standard key or OIDC request falls back to the team’s ordinary AI Gateway credit balance. A Token Pass-scoped key instead stops until the next billing period. The scoped credential is therefore the relevant option when stopping is preferable to drawing down another balance. It is not a guarantee that unrelated applications or credentials cannot incur charges.

Before an unattended run, identify which key the agent actually uses, which balance it can reach and how it reports an exhausted allowance. A setting chosen in one dashboard does not prove an old credential was replaced in every application. Test the intended failure path before relying on it.

Retention settings also matter on fallback

Vercel says requests served through the Token Pass provider have zero data retention. Its routing announcement distinguishes those requests from fallback traffic, which follows the team’s settings. With team-wide zero data retention enabled, fallback routes are restricted to providers offering that property. Do not extend the included route’s data-handling claim to every request made with the same standard key.

The billing guide prorates both the initial charge and credit when a team joins partway through its Pro billing period. Cancellation takes effect at period end, with no refund for that period. The remaining allowance stays usable until then. All team members share the credit, rather than receiving separate $150 balances.

The headline credit is not the same as realized savings

The allowance resets each billing period and unused credit does not roll over. For a team already paying for Pro, the useful comparison is eligible usage it would otherwise purchase, not the largest number on the offer. A team using only a small part of the allowance should not describe the unused remainder as money saved.

The Kimi K3.1 Pro integration announcement lists coding clients including fx, Pi, Cline and Command Code, alongside agents built with the AI SDK. Client compatibility does not settle model suitability. Compare completed changes, test results and total billed usage on the repositories the team actually maintains.

Teams can choose between continuity and a predictable stop, but should make that choice explicitly. Keep the subscription charge, underlying plan and any ordinary Gateway usage separate in the budget. For a related distinction between routing policy and actual charges, see our OpenRouter Auto Code cost and routing guide.

AI-generated editorial illustration, not a product interface. Vercel, the Vercel design, Next.js and related marks, designs and logos are trademarks or registered trademarks of Vercel, Inc. or its affiliates in the US and other countries.

Jordan Reid
Jordan Reid

Jordan Reid is focused on AI tools, agents, developer products, and the way technology changes everyday work. Jordan approaches a launch from the user’s side of the screen. What can it actually help someone finish? The voice is practical, conversational, and skeptical of products that turn a simple job into five new settings. Coverage follows coding assistants, creative software, browser agents, and the workflows around them, with attention to pricing, permissions, setup, and the human work that remains.

Leave a Reply

Your email address will not be published. Required fields are marked *

Gravatar profile