Looking for an alternative to Portkey?
Portkey, now Palo Alto Networks' Prisma AIRS AI Gateway, is an enterprise control plane for 1,600+ models priced by logs, with usage billed by your own providers; Standard Compute includes the compute in one flat plan per key.
TL;DR
Stay on Portkey if you need VPC data planes, HIPAA or a Palo Alto security rollout. Switch if you want coding-agent compute and its bill in one plan per key.
Portkey is now Prisma AIRS AI Gateway
Palo Alto Networks has not published Prisma AIRS AI Gateway prices; on October 13, 2026 portkey.ai still listed its own plans under a "Portkey is now PRISMA AIRS AI Gateway" banner.
| Date | What happened |
|---|---|
| Apr 30, 2026 | Palo Alto Networks announces it will acquire Portkey |
| May 29, 2026 | The acquisition closes |
| Jul 16, 2026 | Prisma AIRS AI Gateway is generally available, "the AI control plane for the enterprise" |
What Portkey offers
Developer is free with 10k recorded logs a month. Production is $49 a month for 100k recorded logs, plus $9 per extra 100k. Enterprise is custom and adds RBAC, SSO, VPC hosting and HIPAA. Model usage is billed by your own providers. (Verified 2026-10-13.)
- Gateway features (routing configs, Messages and Responses APIs) are on every plan, including the free one
- Documented setups for Claude Code, Codex, OpenCode and Cline, with any model in its catalog
- Conditional routing, load balancing and fallbacks you control in a config
- SOC 2, ISO 27001, GDPR and HIPAA, with a hybrid deployment that keeps the data plane in your VPC on Enterprise
- Now backed by Palo Alto Networks' security suite
Why people look for an alternative
- You still pay every provider per token; Portkey's plans price the logs, not the inference
- Budget limits on API keys and workspaces are for Enterprise and "select Pro customers", and a provider budget cannot be edited once set
- Routing is configs you write: a cost-saving split is a fixed weight you choose, not a model picked per request
- Direction is now set by an enterprise security vendor
Which one fits your work
| Pick | When |
|---|---|
| Standard Compute | One bill with compute included: a flat plan per key instead of log fees plus provider invoices |
| The router picks an efficient model for each request, so you don't maintain weights and fallback configs | |
| A plan or limit on every key, on every plan, without an Enterprise contract | |
| Claude Code on the Anthropic Messages API and Codex on the Responses API, with Europe or US region rules per key, team or company | |
| Portkey | You need guardrails, PII checks, SSO, SCIM and audit logs as part of a Palo Alto security rollout |
| Prompts must stay in your own VPC: Portkey's hybrid deployment keeps the data plane there | |
| You want to use existing provider contracts, cloud credits or negotiated rates | |
| You route every LLM call in the company through one governed gateway |
Switching takes one config change
Standard Compute is OpenAI-compatible, so any tool or SDK that lets you set a custom base URL migrates in minutes:
Base URL = https://api.stdcmpt.com/v1 API key = your Standard Compute key Model = standardcompute
Setup guides for OpenClaw, Hermes, OpenCode, Cursor, Cline, Aider and more are on the integrations page.
FAQ
Who owns Portkey now?
Palo Alto Networks, since the acquisition closed in May 2026. It sells Portkey as Prisma AIRS AI Gateway.
Does Portkey sell model usage?
No. You add your own provider credentials, so tokens are billed by OpenAI, Anthropic or whichever provider you connect, and Portkey's plans price logs and features. Standard Compute includes the compute in the plan price.
Can I run Claude Code through Portkey?
Yes. Portkey documents ANTHROPIC_BASE_URL=https://api.portkey.ai and its /v1/messages endpoint on every plan. On Standard Compute you set ANTHROPIC_BASE_URL to https://api.stdcmpt.com, and the same key serves Codex through /v1/responses and other agents through https://api.stdcmpt.com/v1.
Is there an open-source Portkey alternative?
Portkey's own gateway is MIT-licensed on GitHub, with no commits on its main branch since May 25, 2026. LiteLLM is the most widely used self-hosted gateway. Both leave you paying providers per token and running the infrastructure; Standard Compute is the hosted option with compute included.
Sources: Palo Alto Networks: acquisition announced; Palo Alto Networks: acquisition completed; Prisma AIRS AI Gateway GA; Portkey-AI/gateway on GitHub; Portkey pricing; Portkey budget limits; Portkey conditional routing; Portkey Claude Code docs; Prisma AIRS AI Gateway
