OpenRouter alternatives: what to use and when
OpenRouter is a good product, and the right alternative depends on what you need to change: the fee, the way you pay, where it runs, or how much you manage yourself. Below is a side by side of six options, built from each vendor's public pricing pages. We run OneProvider, so we say plainly where it fits and where it does not.
What OpenRouter offers
OpenRouter is a hosted router with one API for 600+ models from 80+ providers. It is OpenAI-compatible, and it can fall back to another provider when one fails.
- Inference is billed at the provider list price. OpenRouter's fee is charged when you buy credits: 5.5% on Standard and 8% on Business.
- Bring your own key: $25,000 of list price inference a month with no OpenRouter fee, then 5%.
- Payment by card, Alipay and USDC crypto.
- Free models are capped at 50 requests a day, or 1,000 a day after you have bought at least $10 of credits.
- For Claude Code it is documented to work only with the Anthropic first-party provider, so set that provider as the top priority.
Source: openrouter.ai pricing and FAQ pages, read on 16 November 2026. Check them for changes before you decide.
Why people look for something else
The reasons we hear are practical, not ideological. Each one points to a different alternative.
- Control: you want to run the gateway yourself, next to your own infrastructure.
- Observability: you need logs, budgets, guardrails and prompt management for a team.
- Fees: you want a different cost structure on credit purchases.
- Payment methods: your card, region or company policy does not fit the available options.
- Claude Code: you want an endpoint that speaks the Anthropic Messages format natively.
Six options side by side
| Option | What it is | How you pay | Worth knowing |
|---|---|---|---|
| OpenRouter | Hosted router, 600+ models | Prepaid credits, 5.5% fee on purchases | Provider list price on every request, built-in fallback |
| LiteLLM | Open-source library and self-hosted gateway | Free software, you pay each provider directly | You run it. Virtual keys, cost tracking and admin UI included |
| Portkey | Gateway with observability and prompt management | Free plan with 10k logs a month, Production $49 a month | Plans are priced on logs and features, not on tokens |
| Vercel AI Gateway | Hosted gateway inside the Vercel platform | Credits, no markup or platform fee on tokens | $5 monthly free credit on the free tier. Card processing fees may apply |
| Cloudflare AI Gateway | Gateway on the Cloudflare platform | Core features free, 5% fee on Unified Billing credits | Credit card only for Unified Billing. Analytics, caching, rate limits |
| OneProvider | Hosted gateway, one key and one USD balance | Crypto, cards of any country with Apple Pay and Google Pay, SBP/QR | Native Anthropic and OpenAI endpoints. Curated catalog, rates published |
How to choose
| If you need | Look at |
|---|---|
| The widest model choice and automatic provider fallback | OpenRouter |
| Full control and self-hosting | LiteLLM |
| Team logs, budgets, guardrails and prompt management | Portkey |
| A gateway inside a Vercel or Cloudflare project you already run | Vercel AI Gateway or Cloudflare AI Gateway |
| A native Anthropic endpoint for Claude Code, one balance across several keys with per-key limits, top-ups by crypto or card | OneProvider |
Be honest about the trade-offs. OneProvider has a curated catalog, not 600+ models, and it does not offer automatic routing of a request across dozens of providers by price. If that routing is what you want, OpenRouter is built for it.
How OneProvider works
You top up one USD balance and issue keys from the cabinet at app.oneprovider.dev or in the Telegram bot @oneprovider_robot. Each key has a name, an optional spend limit and a term of up to 90 days. Three keys are available by default. The balance does not expire and there is no subscription.
Fees depend on how you pay and are shown before the payment. Crypto carries a 3% provider fee plus the network fee. Servers are in Frankfurt and traffic uses TLS 1.3.
- Claude Sonnet 5.5: Input / 1M tokens $2.00, Output / 1M tokens $10.00, Context 1M
- Claude Opus 5.5: Input / 1M tokens $4.00, Output / 1M tokens $20.00, Context 1M
- GPT-5.5: Input / 1M tokens $5.00, Output / 1M tokens $30.00, Context 1050K
- DeepSeek V4 Flash: Input / 1M tokens $0.30, Output / 1M tokens $1.20, Context 1M
Per-token rates come from the public catalog, so you can compare them with any other gateway.
Switching takes two lines
The OpenAI-compatible endpoint is https://api.oneprovider.dev/v1. If your code already uses the OpenAI SDK, change the base URL and the key.
from openai import OpenAI
client = OpenAI(
base_url="https://api.oneprovider.dev/v1",
api_key="YOUR_API_KEY",
)
r = client.chat.completions.create(
model="gpt-5.5",
messages=[{"role": "user", "content": "Say hi"}],
)
print(r.choices[0].message.content)For Claude Code, use the Anthropic endpoint without /v1.
export ANTHROPIC_BASE_URL=https://api.oneprovider.dev
export ANTHROPIC_AUTH_TOKEN=YOUR_API_KEYFrequently asked questions
What is the best alternative to OpenRouter?
It depends on the need. LiteLLM suits self-hosting, Portkey suits team observability, Vercel and Cloudflare suit projects already on those platforms, and OneProvider suits a native Anthropic endpoint with one prepaid balance.
Is there an OpenRouter alternative with no platform fee?
Vercel states that AI Gateway adds no markup and no platform fee on tokens, though payment processing fees may apply. LiteLLM is free software, but you pay each provider directly and run the gateway yourself.
Does OpenRouter work with Claude Code?
Yes, through https://openrouter.ai/api. OpenRouter documents that it is only guaranteed with the Anthropic first-party provider, and recommends making that provider the top priority.
Can I pay for an LLM gateway with crypto?
OpenRouter accepts USDC and OneProvider accepts crypto top-ups. Cloudflare Unified Billing accepts credit cards only. Check each vendor's payment page for your region.
Can I self-host an OpenRouter alternative?
Yes. LiteLLM runs as a container with virtual keys, cost tracking and an admin UI. You then manage uptime, updates and provider keys yourself.
Is OneProvider compatible with the OpenAI SDK?
Yes. Use https://api.oneprovider.dev/v1 as the base URL with your key. The Anthropic-native endpoint is https://api.oneprovider.dev, without /v1.
Claude, ChatGPT, GPT, Cursor and other product names are trademarks of their owners and are used only to describe compatibility. We do not give legal advice: following local rules stays with the user.