Two ways to connect, eight clients covered

Pick your client

The two endpoints

The two paths differ because the clients append different things. An OpenAI client appends /chat/completions, which is why that base already carries /v1. An Anthropic client appends /v1/messages, so its base stops at /api/gateway. Pasting one into the other's setting gives you a /v1/v1/... path and a 404 — the most common mistake setting this up.

Getting a key

Both routes need one. Sign in to the console, open My → API Keys, and create one. Keys start with sk-.

You can read the model list without a key at all; /v1/models is public, which is worth doing before you sign up if you want to check coverage first.

The ten-second version

# 1. is the gateway up? no key needed: this checks reachability, not auth
curl -s https://freemodel.online/v1/models | head -c 120

# 2. a real call, which does check your key
curl -X POST https://freemodel.online/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-your-key" \
  -d '{"model":"fm/qwen-plus","messages":[{"role":"user","content":"Hello"}]}'

When something does not work

When not to use FreeModel

Nothing below is a limitation you can configure around. If any of them applies, a different gateway is the right answer: see the alternatives.

Beyond chat

Six capabilities, one alias each, all callable from the same key and the same base URL. They are grouped separately from the chat models because they share a rule: an alias only moves to a different model when the current one has been retired — never on a timeout, and never silently.

Next

Set up in three steps

Every client on this page follows the same three steps. What changes between them is only where the value goes.

1. Step 1: Create a key in the console under My → API Keys. It starts with sk-. Nothing else here works without one, though the model list is readable without a key.
2. Step 2: Point the client at the base URL that matches its API shape. OpenAI-shaped clients take https://freemodel.online/v1; Anthropic-shaped clients take https://freemodel.online/api/gateway. The client appends its own path, so the base never contains /chat/completions or /v1/messages.
3. Step 3: Name a model or a route wherever the client normally expects a model name. A concrete id from /v1/models always works; the auto/* modality routes and the three tier aliases — fm-v1-lite, fm-v1-standard, fm-v1-pro — hand the choice to the router instead.

What you point at is a catalogue rather than one provider. The 199 listed models put chat, image, video, speech synthesis, speech recognition, embeddings and rerank in one list; 77% of them are chat models and 23% are everything else.[1] Six auto/* routes cover the non-chat side: auto/image, auto/video, auto/tts, auto/asr, auto/embed and auto/rerank. Each carries a candidates field, so you can read how many models sit behind a route before you rely on it. Chat is reached by naming a concrete id, or one of three subscription tier aliases: fm-v1-lite, fm-v1-standard and fm-v1-pro. One key covers all of it, in both the OpenAI and Anthropic API shapes.

If you only want to check coverage, do that before signing up: the list endpoint answers anonymously.

Docs FAQ

A few questions come up whatever client you picked. One key covers all of them: 199 listed models, 77% of them chat models and the rest spread across image, video, speech and embeddings[1], so the friction is almost always about endpoints rather than about credentials.

Sources

Every third-party figure quoted on this site was read from the vendor documentation linked below, on the date stated. None of these links are affiliate links and none of these vendors paid to be listed. If a number here disagrees with the page it came from, the page is right — tell us and it gets corrected.

• OpenRouter — credit-purchase fees and free-model rate limits
• OpenRouter — rate-limit reference
• OpenRouter — public model list
• LiteLLM — licence, deployment options and supported models
• Cloudflare AI Gateway — features and pricing
• Vercel AI Gateway — features
• Portkey — plans
• Helicone — plans
• Model Context Protocol — specification
• OpenAI Python SDK — base_url and OPENAI_BASE_URL
• Anthropic Messages API — request format

Where the percentages come from

Every share quoted on this site is counted from the live model list, not estimated. Divide the two counts in the last column, round to the nearest whole percent, and you get the number in the first.

[1] Counted on 29 September 2026 from GET https://freemodel.online/v1/models, which needs no key and no account. Every row uses all 199 listed entries as the denominator, because the only thing the endpoint states for every entry is its modality. It no longer publishes per-model capability flags such as tool calling or reasoning, so this site quotes no shares for them. Fetch the list and recount whenever you want to check the arithmetic.