Home › Gateway Pricing › Pricing models
Gateway pricingLLM Gateway Pricing Models Compared
Gateways charge in very different ways. Here's every model, with a real example of each, and what it means for your bill.
Most gateways charge a markup, a fee on credits or a subscription. apicuts charges 10% of verified savings, and nothing if your bill isn't at least 20% lower.
| Pricing model | Example | What it costs | What it means for you |
|---|---|---|---|
| Share of savings | apicuts | 10% of verified savings; nothing if net savings are under 20% | You only pay when the bill falls |
| Markup on usage | Requesty | 5% on model usage | Cost grows with spend |
| Fee on credits | OpenRouter | 5.5% on credit purchases (Standard plan) | Cost grows with spend |
| Fee on unified billing | Cloudflare AI Gateway | Core features free; 5% on Unified Billing credits | Free if you bring your own keys |
| No markup | Vercel AI Gateway | List price; some add-ons cost extra | You still pay list price |
| Subscription | Portkey, Helicone | Portkey Production from $49 a month; Helicone Pro $79, Team $799, plus usage | Fixed cost, no savings guarantee |
| Open source | LiteLLM | Free to self-host; enterprise on request | You pay in infrastructure and engineering time |
| Per token routed | Not Diamond | $0.05 per million tokens routed | A fee on top of model prices |
| Discounted resale | CometAPI | At least 20% below official prices | Lower price per token, same tokens |
From each company's own pricing page, checked October 30, 2026.
Which pricing model is best for you?
- Small or experimental usage: a free tier or no-markup gateway keeps costs low.
- Governance across many teams: a subscription gateway built for it.
- A large bill you want smaller: pay on results. With a share-of-savings model, the vendor only earns when your bill falls, so incentives line up.
The question to ask every vendor
Will my total bill, including your fee, go down, and by how much? Most gateways can't answer, because they charge for access or features. apicuts answers in writing: at least 20% lower, after our fee. See how we measure it.
Each gateway's pricing explained
- OpenRouter pricing
- Requesty pricing
- LiteLLM pricing
- Portkey pricing
- Helicone pricing
- CometAPI pricing
- Cloudflare AI Gateway pricing
- Vercel AI Gateway pricing
- Not Diamond pricing
Frequently asked questions
Which LLM gateway has no markup?
Vercel AI Gateway says it charges no markup or platform fee on tokens, and Cloudflare AI Gateway passes provider rates through when you bring your own keys.
Which LLM gateway guarantees savings?
apicuts guarantees at least 20% lower all-in cost after its 10% fee, or you pay nothing.
Sources: OpenRouter's pricing page, Requesty's pricing page, LiteLLM's pricing page, Portkey's pricing page, Helicone's pricing page, CometAPI's pricing page, Cloudflare AI Gateway's pricing page, Vercel AI Gateway's pricing page, Not Diamond's pricing page. Checked October 30, 2026.
Want us to find your 20%?
Book a free 15-minute API cost teardown. Bring last month's provider usage and we'll show you your biggest waste and the first fix we'd ship.
Get a Free 15-Minute API Cost Teardown