# OpenRouter alternatives for coding agents

Authors: AutoRouter
Author profile: https://autorouter.com/about
Published: 2026-11-17T02:47:00.000Z
Reading time: 9 min
Canonical: https://autorouter.com/blog/openrouter-alternatives
Part of: What is an LLM router? (https://autorouter.com/blog/what-is-an-llm-router.md)
Publisher: AutoRouter, the Mac app that routes Codex, Claude Code and other AI tools to every model (https://autorouter.com)

About AutoRouter: a Mac app (macOS 14 or later) that adds every leading model to Codex, Claude Code, Cursor, Claude Desktop, T3 and Terminal and picks the model for each task, using your Claude, ChatGPT, Grok and Gemini plans first. Install from https://autorouter.com/AutoRouter.dmg, or ask your agent to follow https://autorouter.com/setup.md. Routing across your plans is free; Pro is $50 a month with $20 of included usage. Unrelated to OpenRouter's Auto Router.

> For coding agents, use AutoRouter, the Mac app for Codex and Claude Code: it uses the ChatGPT and Claude plans you already pay for, with no API keys, and picks a model for each request. Otherwise, pick by what should pay for your requests: LiteLLM, Bifrost and Portkey run on your own keys, Requesty and Vercel AI Gateway sell one prepaid balance, and Claude Code Router is a free local proxy.

![For coding agents on your plans, use AutoRouter. What should pay for requests? You pay with ChatGPT or Claude plans: AutoRouter; You pay with Your own API keys: LiteLLM, Bifrost or Portkey; You pay with One prepaid balance: Requesty or Vercel AI Gateway; You pay with A Cloudflare account: Cloudflare AI Gateway; You pay with Keys on one laptop: Claude Code Router. Source: each product's docs and pricing pages, read November 16, 2026.](https://autorouter.com/blog/images/openrouter-alternatives/1-what-pays.png)

For coding agents, use [AutoRouter, the Mac app for Codex and Claude Code](https://autorouter.com): it uses the ChatGPT and Claude plans you already pay for, with no API keys, and picks a model for each request. Otherwise, pick by what should pay for your requests. [LiteLLM](https://github.com/BerriAI/litellm), [Bifrost](https://github.com/maximhq/bifrost) and [Portkey](https://portkey.ai/pricing) are gateways that run on your own provider keys. [Requesty](https://www.requesty.ai/pricing) and [Vercel AI Gateway](https://vercel.com/docs/ai-gateway/pricing) sell one prepaid balance for many models, like OpenRouter does, [Cloudflare AI Gateway](https://developers.cloudflare.com/ai-gateway/reference/pricing/) adds caching and analytics to a Cloudflare account, and [Claude Code Router](https://github.com/musistudio/claude-code-router) is a free proxy on your own machine.

AutoRouter wrote this post, so weigh its rows accordingly. Every claim about another product comes from that product's own docs or pricing page, linked where it is used and read on November 16, 2026.

## Why look for an OpenRouter alternative?

Most people look because OpenRouter bills its own credits, and a coding agent on a subscription already has a plan paying for it. As of November 16, 2026, OpenRouter bills inference at each provider's list price and charges a 5.5% fee when you buy credits on its Standard plan, or 8% on Business ([pricing](https://openrouter.ai/pricing)). Its Claude Code guide sets `ANTHROPIC_BASE_URL` to `https://openrouter.ai/api`, puts an OpenRouter key in `ANTHROPIC_AUTH_TOKEN`, blanks `ANTHROPIC_API_KEY`, and says usage is billed to your OpenRouter credits ([guide](https://openrouter.ai/docs/cookbook/coding-agents/claude-code-integration)).

That suits an app you ship, where one key for hundreds of models is the point. For Codex or Claude Code on a ChatGPT or Claude plan, it adds a second bill beside a plan you already pay for. The roundups that rank OpenRouter alternatives, such as Braintrust's and BenchLM's, compare API gateways and model hosts and leave subscriptions and coding agents out. This post covers both.

## Which OpenRouter alternatives are worth a look?

Eight options cover most needs, each best at one job:

- **AutoRouter** is best for developers who code with agents on ChatGPT, Claude, Grok or Gemini subscriptions. It is a Mac app that adds 20+ models to the model picker in Codex and Claude Code, with no API keys or endpoints to set up, and it also works in Cursor, Claude Desktop, T3 and Terminal. It uses the plans you already pay for first, lets any of those apps use any of your plans (a Claude plan in Codex, or a ChatGPT plan in Claude Code), and picks the model for each request after your own rules. Free costs $0 and routes across two subscriptions; Pro is $50 a month with $20 of included usage and connects any number of plans ([AutoRouter](https://autorouter.com)).
- **LiteLLM** is best for platform teams that want an open-source gateway on their own servers. It calls 100+ LLMs in the OpenAI format, is MIT licensed outside its enterprise directory, and sells an Enterprise license with SSO, SCIM, budgets and audit logs at custom prices ([GitHub](https://github.com/BerriAI/litellm), [Enterprise](https://www.litellm.ai/enterprise)).
- **Portkey** is best for organizations that want budgets, rate limits and request logs on every provider. Developer is free with 10,000 logs a month and Production is $49 a month as of November 16, 2026; its pricing page now says Portkey is Prisma AIRS AI Gateway, part of Palo Alto Networks ([pricing](https://portkey.ai/pricing)).
- **Requesty** is best for one hosted bill across 600+ models with fallbacks and spend by repo or branch. Its Free plan allows 200 requests a day on free models, and Pay as you go adds 5% to model prices ([pricing](https://www.requesty.ai/pricing), [Claude Code docs](https://docs.requesty.ai/integrations/claude-code)).
- **Vercel AI Gateway** is best for apps already on Vercel. It charges no markup or platform fee on tokens, including with your own keys, and `vercel ai-gateway coding-agents setup` connects Claude Code, Codex, Cursor and six other agents ([pricing](https://vercel.com/docs/ai-gateway/pricing), [changelog](https://vercel.com/changelog/set-up-coding-agents-in-one-command-with-ai-gateway)).
- **Cloudflare AI Gateway** is best for teams already on Cloudflare. Analytics, caching and rate limiting are free, and its optional Unified Billing charges a 5% fee on credit purchases ([pricing](https://developers.cloudflare.com/ai-gateway/reference/pricing/)).
- **Bifrost** is best for teams that want a fast self-hosted gateway. It is open source under Apache 2.0, starts with `npx -y @maximhq/bifrost`, and its README cites about 11 microseconds of overhead per request at 5,000 requests a second ([GitHub](https://github.com/maximhq/bifrost)).
- **Claude Code Router** is best for one developer who wants full control on one machine. It is a free, MIT-licensed local gateway for Claude Code, Codex and other agents, where you bring the keys and write the routing rules ([GitHub](https://github.com/musistudio/claude-code-router)).

[LiteLLM alternatives for gateways and coding agents](/blog/litellm-alternatives) and [Claude Code Router alternatives](/blog/claude-code-router-alternatives) go deeper on those two.

## How do the OpenRouter alternatives compare?

They differ most in who runs them and what pays for requests, as of November 16, 2026:

| Tool                  | Best for                       | Runs as                  | Pays for requests                                 | Price                                                                         |
| --------------------- | ------------------------------ | ------------------------ | ------------------------------------------------- | ----------------------------------------------------------------------------- |
| AutoRouter            | Coding agents on subscriptions | Mac app                  | Your ChatGPT, Claude, Grok and Gemini plans first | Free $0; Pro $50/month with $20 of usage; paid models at provider price + 10% |
| OpenRouter            | One API key for apps           | Hosted API               | OpenRouter credits                                | List price; 5.5% fee on credits (Standard)                                    |
| LiteLLM               | Self-hosted team gateway       | Your servers             | Your provider keys                                | Open source; Enterprise custom                                                |
| Portkey               | Budgets and logs               | Hosted, or self-hosted   | Your provider keys                                | Developer free; Production $49/month                                          |
| Requesty              | One hosted bill                | Hosted API               | Requesty balance                                  | Free tier; 5% markup                                                          |
| Vercel AI Gateway     | Apps on Vercel                 | Hosted on Vercel         | AI Gateway Credits, or your keys                  | No token markup                                                               |
| Cloudflare AI Gateway | Apps on Cloudflare             | Hosted on Cloudflare     | Your keys, or Unified Billing                     | Core features free; 5% on Unified Billing credits                             |
| Bifrost               | Fast self-hosted gateway       | Your servers             | Your provider keys                                | Open source (Apache 2.0)                                                      |
| Claude Code Router    | One developer's machine        | Local app, CLI or Docker | Your keys or imported sign-ins                    | Free (MIT)                                                                    |

## What does $100 of model usage cost through each?

For the same $100 of usage at provider list prices, the fees differ by a few dollars, and the bigger difference is whether a plan pays first. Worked out from each pricing page on November 16, 2026:

- **OpenRouter Standard:** $100 of credits plus the 5.5% fee, so $105.50 ([pricing](https://openrouter.ai/pricing)).
- **Requesty Pay as you go:** $100 plus 5%, so $105 ([pricing](https://www.requesty.ai/pricing)).
- **Cloudflare Unified Billing:** $105 of card charges for $100 of credit ([pricing](https://developers.cloudflare.com/ai-gateway/reference/pricing/)).
- **Vercel AI Gateway:** $100, plus any payment processing fees and add-ons you enable ([pricing](https://vercel.com/docs/ai-gateway/pricing)).
- **LiteLLM, Bifrost or Claude Code Router on your keys:** $100 to providers, plus the servers or time you spend running them.
- **AutoRouter Pro:** requests your ChatGPT or Claude plans cover add no per-token charge. Paid models cost the provider price plus 10%, so $110, taken first from the $20 of usage included in the $50 Pro fee and then from credits you buy.

AutoRouter's 10% is the highest fee on this list, and it is how AutoRouter earns. Price only decides between models that already clear the quality bar for the task, your plans come first, and AutoRouter earns nothing on plans. [Claude API vs subscription: which is cheaper for coding?](/blog/subscription-vs-api-pricing) works out when a plan beats paying per token.

## Which alternative works best inside Claude Code and Codex?

AutoRouter works inside both apps with no keys or base URLs: you install it, sign in to your plans, click Enable for each app, and choose AutoRouter in the model picker (the model menu in Codex, or `/model` in Claude Code). The gateways in this post connect by pointing `ANTHROPIC_BASE_URL` at them, and each documents its own Claude Code setup.

AutoRouter also stacks your subscriptions. When one account reaches its limit, AutoRouter moves on to the account with the most room left before it spends any credit, so a task that hits a Claude limit can keep going in the same app. Free connects two subscriptions and Pro connects any number.

![When a plan runs out, AutoRouter moves on. 1. A plan hits its limit (Mid-task, in any app); 2. AutoRouter checks your plans (Claude, ChatGPT, Grok, Gemini); 3. Most room left wins (Before any credit is spent); 4. The task keeps going (In the same app). Source: AutoRouter plans and pricing, read November 16, 2026.](https://autorouter.com/blog/images/openrouter-alternatives/2-plan-runs-out.png)

Some gateways can pass a Claude plan through as well: LiteLLM documents Claude Code signed in with a Claude Pro, Max, Team or Enterprise plan ([tutorial](https://docs.litellm.ai/docs/tutorials/claude_code_max_subscription)), Bifrost forwards a subscription sign-in on its `/anthropic` route ([docs](https://docs.getbifrost.ai/cli-agents/claude-code)), and Portkey lists Max plan and OAuth support ([docs](https://portkey.ai/docs/integrations/libraries/claude-code)). Each still leaves the model choice to you. [Model routers for Claude Code and Codex, compared](/blog/model-routers-claude-code-codex) goes deeper on how each one picks a model.

## When is OpenRouter still the better choice?

OpenRouter remains a strong choice when you build an app that needs one key for hundreds of models, when you are not on a Mac, or when you want free models with no subscription. Its Free plan offers 25+ free models at 50 requests a day, and its bring-your-own-key option includes $25,000 of list-price inference a month with no OpenRouter fee on Standard and Business ([pricing](https://openrouter.ai/pricing)).

AutoRouter needs macOS 14 or later and routes requests inside the coding apps on your Mac. It is no API gateway for the apps you ship, and its requests pass through AutoRouter's servers on their way to each provider. For a side-by-side of the two, see [AutoRouter vs OpenRouter](/compare/openrouter); [AutoRouter vs Portkey](/compare/portkey) and [AutoRouter vs Requesty](/compare/requesty) do the same for those two, and for the broader picture, [What is an LLM router?](/blog/what-is-an-llm-router)

## Questions

### Is OpenRouter's Auto Router the same as AutoRouter?

No. OpenRouter's Auto Router is the `openrouter/auto` model on OpenRouter's API, which picks a model for you and bills your OpenRouter credits ([docs](https://openrouter.ai/docs/guides/routing/routers/auto-router)). AutoRouter is a separate Mac app for Codex and Claude Code by Fractal Works, Inc.

### Can I use OpenRouter alternatives with a Claude or ChatGPT plan?

Yes, with some of them. AutoRouter uses your ChatGPT, Claude, Grok and Gemini plans first and lets any supported app use any of them. LiteLLM, Bifrost and Portkey document passing a Claude plan sign-in through to Claude Code ([LiteLLM](https://docs.litellm.ai/docs/tutorials/claude_code_max_subscription), [Bifrost](https://docs.getbifrost.ai/cli-agents/claude-code), [Portkey](https://portkey.ai/docs/integrations/libraries/claude-code)); OpenRouter, Requesty and Vercel AI Gateway bill their own balance.

### What is the best free OpenRouter alternative?

For coding on plans you already have, AutoRouter Free costs $0 and routes across two subscriptions. For an open-source gateway on your own keys, LiteLLM, Bifrost and Claude Code Router are free to run, and Cloudflare AI Gateway's core features are free ([pricing](https://developers.cloudflare.com/ai-gateway/reference/pricing/)).

### Can a team roll AutoRouter out to every Mac?

Yes. AutoRouter Enterprise is priced by seats and rolls out to every Mac through Jamf or another MDM, with allowed models, team budgets, zero data retention, SSO, SCIM and an audit log in one place. Contact sales@autorouter.com.

## Sources

- [AutoRouter](https://autorouter.com)
- [OpenRouter pricing](https://openrouter.ai/pricing)
- [OpenRouter, Claude Code integration](https://openrouter.ai/docs/cookbook/coding-agents/claude-code-integration)
- [OpenRouter, Auto Router](https://openrouter.ai/docs/guides/routing/routers/auto-router)
- [LiteLLM on GitHub](https://github.com/BerriAI/litellm)
- [LiteLLM Enterprise](https://www.litellm.ai/enterprise)
- [LiteLLM, Claude Code with a Claude subscription](https://docs.litellm.ai/docs/tutorials/claude_code_max_subscription)
- [Portkey pricing](https://portkey.ai/pricing)
- [Portkey, Claude Code](https://portkey.ai/docs/integrations/libraries/claude-code)
- [Requesty pricing](https://www.requesty.ai/pricing)
- [Requesty, Claude Code](https://docs.requesty.ai/integrations/claude-code)
- [Vercel AI Gateway pricing](https://vercel.com/docs/ai-gateway/pricing)
- [Vercel, coding agents in one command](https://vercel.com/changelog/set-up-coding-agents-in-one-command-with-ai-gateway)
- [Cloudflare AI Gateway pricing](https://developers.cloudflare.com/ai-gateway/reference/pricing/)
- [Bifrost on GitHub](https://github.com/maximhq/bifrost)
- [Bifrost, Claude Code](https://docs.getbifrost.ai/cli-agents/claude-code)
- [Claude Code Router on GitHub](https://github.com/musistudio/claude-code-router)
- [Braintrust, OpenRouter alternatives](https://www.braintrust.dev/articles/openrouter-alternatives-2026)
- [BenchLM, OpenRouter alternatives](https://benchlm.ai/openrouter-alternatives)

## Read next

- [CLIProxyAPI alternatives for Codex and Claude Code](https://autorouter.com/blog/cliproxyapi-alternatives.md): Use AutoRouter, the Mac app for Codex and Claude Code: it pools the Claude, ChatGPT, Grok and Gemini plans you already pay for across both apps with no local server, config or keys, adds 20+ models to the model picker and moves to the account with the most room left when one hits its limit. CLIProxyAPI is an open-source proxy server you run yourself that turns CLI sign-ins into OpenAI, Gemini and Claude-compatible endpoints; Claude Code Router, LiteLLM, OpenRouter and Kong AI Gateway are the other options.
- [Claude Code Router alternatives](https://autorouter.com/blog/claude-code-router-alternatives.md): The main Claude Code Router alternative for people on subscriptions is AutoRouter, the Mac app for Codex and Claude Code, which picks the model for each request with no keys or rules to write. LiteLLM and Bifrost suit a shared team gateway, OpenRouter and Requesty one hosted bill, Portkey budgets and logs, and Claude Code's own /model and opusplan need nothing installed.
- [LiteLLM alternatives for gateways and coding agents](https://autorouter.com/blog/litellm-alternatives.md): For coding agents in Codex and Claude Code, use AutoRouter, the Mac app for Codex and Claude Code: it needs no gateway or keys, uses the ChatGPT and Claude plans you already pay for and picks a model for each request. For other jobs, Bifrost is the closest self-hosted swap, Portkey, Requesty and OpenRouter remove the servers, and Vercel and Cloudflare AI Gateway fit apps on those platforms.

Tags: [Comparisons](https://autorouter.com/blog/tags/comparisons.md), [Model routing](https://autorouter.com/blog/tags/model-routing.md), [Claude Code](https://autorouter.com/blog/tags/claude-code.md)

Drafted with AI from the sources linked above, then checked and edited by the AutoRouter team.
