Category
5 of 20 products we track
A cloud platform is a hyperscaler surface that offers models from several vendors under one contract, one bill and one identity system. You are not adding a party to your architecture so much as using more of a provider you already have, which is why these tend to arrive through procurement rather than through engineering.
The reason to choose this category is rarely the product and almost always the paperwork. The contract is signed, the security review is done, the committed spend is already on the books, and access is governed by the same cloud IAM roles as everything else you run. For a large organisation that combination outweighs a great deal of feature difference.
What you give up is reach and speed. A hyperscaler catalogue is a curated subset of what exists, negotiated vendor by vendor, so a model available everywhere else on launch day may take months to appear or never arrive. If your product depends on using the newest model quickly, this is the wrong category to depend on exclusively.
These products generally do not charge a gateway fee at all — you pay per token at the platform rate, and the routing layer is included. That makes the headline look free next to a marketplace percentage, which is misleading in both directions: the per-token rate itself may be higher or lower than elsewhere, and committed-spend discounts you already hold can move it substantially. Compare the delivered token price, not the fee.
The pricing guide works through all four charging mechanisms — token markupMarkup: A percentage the gateway adds on top of what the model actually costs. Some charge none at all and make money elsewhere., credit feesCredit or top-up fee: A cut taken when you add money to a prepaid balance, typically around 5%. Easy to miss because it is not a markup on tokens — but you pay it on every dollar you load., per-seat, and self-hostedSelf-hosted: You run the software on your own infrastructure. No third party sees your traffic, and there is no vendor fee — but you own the uptime, the patching, and the upgrades. — with figures computed at four real workloads, and the cost estimator runs the same model against your own volumes.
If you are not already committed to that cloud, this is the most expensive way to enter it. The advantages here are almost entirely advantages of incumbency.
AWS-managed service for calling foundation models from 19 model providers through one AWS API.
US company
Best for Teams already standardized on AWS that need many model vendors behind one IAM-governed, compliance-attested API.
Microsoft's Azure platform for deploying models from its own and partner catalogs, now branded Microsoft Foundry.
US company · EU region available
Best for Microsoft-centric enterprises that want first-party OpenAI models plus a very large partner catalog under Azure governance.
Edge proxy in front of a curated set of AI providers, with caching, rate limiting, DLP and analytics.
US company
Best for Teams already on Cloudflare Workers who want free caching, analytics, spend limits, DLP and guardrails at the edge.
Google Cloud's model platform for Gemini plus 200+ Model Garden models, now branded Gemini Enterprise Agent Platform.
US company · EU region available
Best for Google Cloud customers who want Gemini alongside third-party models with strong, explicitly documented EU residency and ZDR controls.
Closed-source enterprise AI gateway sold on request tiers, deployable as SaaS or inside the customer's own cloud.
India company
Best for Enterprises that want a fully managed or in-VPC AI gateway with guardrails, MCP governance and SSO/RBAC, and are comfortable with closed source.
Products are sorted by what they actually are, not by what they are marketed as. If none of the above is the shape of your problem, one of these probably is.
It does the job of one — a single authenticated endpoint reaching multiple model vendors with logging and spend controls attached — so it is tracked here. It is kept in a separate category because the purchase, the compliance posture and the pricing shape are all different from a standalone gateway, and mixing them in one ranking would mislead.
Typically no separate fee: the routing and governance layer is bundled and you pay the platform per token. That is why these products show no markup and no seat cost in the catalogue, and why comparing them to a percentage-fee marketplace on fee alone produces a meaningless answer.
Only those the hyperscaler has negotiated into its own catalogue. Coverage is a curated list rather than an open marketplace, so model count and which specific vendors are represented are the fields worth checking before committing.