# What developers actually route through AnyRouter

> The catalog has hundreds of models across many providers, but that's not what traffic looks like. A handful of models absorb almost all of the free-tier volume, model counts skew hard toward a few owner organizations, and most requests never mention a provider name at all. They arrive tagged by which coding agent sent them. Here's the shape of the catalog and the shape of what people actually call.

- Date: July 12, 2026
- Category: data
- HTML: https://anyrouter.dev/blog/what-developers-actually-route

## The catalog is wide, not deep

![The AnyRouter leaderboard: top 30 models by Artificial Analysis intelligence index, with reasoning and math comparisons.](https://cdn.anyrouter.dev/screenshots/leaderboard-with-bg-2026-11.webp)

_The live leaderboard ranks the catalog by published benchmark scores, useful for picking a model, not a claim about our own traffic._

AnyRouter's catalog covers hundreds of models across many providers, addressed by a single `provider/model` string: `anthropic/claude-opus-4-8`, `z-ai/glm-5.2`, `google/gemini-2.5-pro`. Models are organized by the organization that owns them, not by which upstream serves them, and that ownership is lopsided: a handful of labs account for most of the catalog, and a long tail of smaller labs each contribute a model or two.

| Owner | Models in catalog |
| --- | --- |
| openai | 32 |
| qwen | 16 |
| meta | 16 |
| google | 14 |
| anthropic | 12 |
| x-ai | 11 |
| venice | 8 |
| wafer | 7 |
| deepseek | 7 |
| nvidia | 6 |

That split matters because a model can also come in two SKU classes on the Cloudflare side alone, hosted first-party (`@cf/owner/model`) or partner-hosted (`owner/model`, no `@cf` prefix), before you even get to a third-party upstream like DeepInfra or a direct provider API. Breadth in the catalog doesn't mean breadth in what gets called.

## What people actually call for free

Calling any model is one request against a single, OpenAI-compatible endpoint:

```bash
curl https://anyrouter.dev/api/v1/chat/completions \
  -H "Authorization: Bearer sk-ar-v1-..." \
  -H "Content-Type: application/json" \
  -d '{"model": "z-ai/glm-5.2", "messages": [{"role": "user", "content": "Hello"}]}'
```

_Same endpoint, same key, any model in the catalog._

But traffic through the free routes, no credit charged, not a BYOK call, doesn't spread evenly across the catalog. Two models account for the overwhelming majority of free requests, all-time through July 11, 2026:

**Free-served requests by model**

- z-ai/glm-5.2: 1,890
- openai/gpt-4o-mini: 1,303
- anyrouter/free: 449
- z-ai/glm-5.1: 138
- google/gemma-4-31b: 66
- openai/gpt-oss-120b: 11

## Two shapes of usage, not one

Request count alone hides what's actually happening. GLM-5.2 leads on requests and dwarfs everything else on tokens served, north of 200M free tokens, which is the signature of a model running real, sustained work, not one-off tests. gpt-4o-mini leads on requests too, but with a fraction of the tokens per call: it's the classifier-and-glue model, cheap enough to sit inside a loop where each call is a sentence or two.

Those are two different jobs wearing the same 'free model' label, and the request count on its own doesn't tell you which one a given model is doing. The tokens-per-request ratio does.

## Coding agents show up as attribution, not as a model name

A meaningful share of production traffic doesn't come from a developer typing a prompt. It comes from a coding agent making the call on their behalf. Claude Code, Codex, and Cursor all speak different dialects against the same catalog:

| Dialect | Endpoint | Used by |
| --- | --- | --- |
| OpenAI chat | /api/v1/chat/completions | Codex, Cursor, any OpenAI-compatible client |
| Anthropic messages | /api/v1/messages | Claude Code and other Claude-native clients |
| OpenAI Responses | /api/v1/responses | Newer OpenAI-compatible agent tooling |

Because every dialect terminates at the same catalog, a request from Claude Code and a request typed by hand can resolve to the same model. The only difference visible in usage analytics is the app attribution header the client sent, not a separate code path.

## The free tier is funded, not subsidized from one balance sheet

The free capacity behind those request counts isn't purchased in bulk. It's a shared key pool. Members donate spare quota from a provider's own free tier, quotas sum into one pool, and donors earn a share back plus a free plan tier. The more capacity gets donated, the more the free-tier numbers above can grow without a paywall behind them.

![AnyRouter free models catalog page listing models callable at $0](https://cdn.anyrouter.dev/screenshots/free-models-with-bg-2026-11.webp)

_Every model in the free-tier table above is callable right now at anyrouter.dev/free._

## Reading the shape correctly

- A wide catalog and a concentrated traffic pattern aren't a contradiction, reach and popularity are different things
- Request count and token count together show what job a model is doing. Count alone only shows how often it is called
- Model ownership in the catalog skews toward a handful of labs, but the smaller entries still get called, just less often
- Free-tier volume scales with donated capacity, not with a fixed budget

See live catalog, traffic, and per-model usage in your own account.

[Open the dashboard](https://anyrouter.dev/dashboard)
