EXHIBIT D — AI PROVIDER PRICING SCHEDULE

1. Purpose

This Pricing Schedule sets the overage rate charged to a Tenant for AI token consumption that runs on Excede platform-managed credentials and exceeds the Included AI Allowance described in MSA Section 2.6 (and, for beta clients, Section 4 of the Beta Client Addendum). It applies to platform-managed usage regardless of which underlying AI model provider a Tenant selects through the Platform’s administrative settings. It does notapply to usage served by a Tenant’s optional customer-supplied (BYO) OpenAI/Anthropic API key or Vertex AI service account (see Section 5).

2. Methodology

• Overage Rate is a flat, single rate per 1,000,000 tokens for the specific provider/model a Tenant selects, billed at the same rate regardless of a Tenant’s actual input/output mix, for billing simplicity and predictability.
• Rates reflect Excede’s underlying provider costs plus a standard platform margin, consistent with MSA Section 2.6. Excede does not publish its underlying cost or margin structure; the table below is the complete and controlling rate schedule for billing purposes when platform-managed credentials serve the request.
• Rates are reviewed periodically and adjusted consistent with the notice procedure in Section 6 below.

3. Current Provider & Model Rates

All rates are in USD per 1,000,000 tokens. Your $85.00 monthly Included AI Allowance is drawn down by platform-managed AI usage; the Overage Rate below applies to platform-managed usage beyond that allowance for the specific model you have selected.

Overage Rate is the amount invoiced to Tenant per 1,000,000 tokens consumed on platform-managed credentials above the Included AI Allowance, for the specific model the Tenant has selected.

4. Fair Use Allowance

• The Included AI Allowance stated in the MSA and any Addendum (currently $85.00/month) is denominated in dollars, not tokens, and is therefore unaffected by a Tenant’s model selection — it simply buys fewer tokens on a more expensive model and more tokens on a cheaper one.
• Tenants who select a premium model (e.g., GPT-5.5, Claude Opus 4.8) on platform-managed credentials should expect their Included AI Allowance to be consumed faster and overage charges, if any, to be materially higher per-token than for lower-cost models.

5. Customer-Supplied API Keys (BYOK)

• Tenants may optionally supply their own OpenAI or Anthropic API keys, or a Vertex AI (GCP) service-account JSON, in Settings → AI Configuration. This is optional; platform-managed credentials remain available when no BYO credential is configured.
• When a Tenant’s BYO credential serves a request, the Tenant pays that provider directly for the tokens consumed. That usage does not draw down the Included AI Allowance and is not invoiced by Excede under this Schedule.
• If a BYO credential fails (invalid, revoked, or provider outage), the Platform may transparently fall back to platform-managed Vertex AI for that call. Fallback usage is metered under this Schedule and draws down the Included AI Allowance.

6. Rate Adjustments

Excede will review this Schedule at least quarterly and upon any material provider price change. Consistent with MSA Section 2.6 and Addendum Section 4, rate changes take effect no earlier than thirty (30) days after written notice (email to Tenant’s designated contact and/or in-Platform notification) and apply prospectively to the next full calendar month of usage. Superseded versions of this Schedule remain available on request for audit purposes.

7. Revision History

• v1.2— 7/30/2026 — Extend optional BYOK to Vertex AI service-account credentials; Exhibit D still meters only platform-served calls.
• v1.1— 7/30/2026 — Clarify that Exhibit D rates and the Included AI Allowance apply to platform-managed credentials only; optional customer-supplied (BYO) OpenAI/Anthropic keys are billed by the provider directly.
• v1.0— 7/7/2026 — Initial publication. Supersedes the flat $1.50 / 1,000,000 token overage rate previously stated in the Beta Client Addendum, which applied only to a single model.