> ## Documentation Index
> Fetch the complete documentation index at: https://docs.openference.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Model Catalog

> How models are selected, curated, and exposed.

# Model Catalog

Openference maintains a curated list of models.

* The public [Models page](https://openference.com/models) shows the current catalog with pricing and descriptions.
* `GET /v1/models` returns the models your key is allowed to use (filtered by key restrictions).
* We regularly evaluate quality and add/remove models.

## Discovering models at runtime

```bash theme={null}
curl https://api.openference.com/v1/models \
  -H "Authorization: Bearer YOUR_API_KEY"
```

The response is an OpenAI-compatible model list — use each entry's `id` as the `model` value in your requests. See [GET /v1/models](/api-reference/models) for the exact shape.

## Context windows

Each entry in `GET /v1/models` advertises the model's maximum **input** context
window (`context_length`) and maximum output tokens (`max_output_tokens`).
The advertised `context_length` is the safe input budget — it's the model's
total context wall minus `max_output_tokens`, with a small \~2% safety buffer
so input + output stays under the upstream's real limit even allowing for
estimation error. Models with a smaller output cap advertise a larger input
budget. Current flagship values:

| Model             | Context window (input) | Max output |
| ----------------- | ---------------------- | ---------- |
| GLM-5.2           | 850,000                | 128,000    |
| DeepSeek-V4-Pro   | 850,000                | 131,072    |
| DeepSeek-V4-Flash | 850,000                | 131,072    |
| MiniMax M3        | 850,000                | 131,072    |
| Auto              | 250,000                | 131,072    |
| Qwen3 Code        | 915,000                | 65,536     |
| Qwen3.7 Plus      | 915,000                | 65,536     |

If your coding client has a context window setting (OpenCode `limit.context`,
Cline/Roo/Kilo "Context Window Size"), set it to the model's `context_length`.
Clients that are left on their default (often 128k) begin compacting
conversations around \~100k tokens, well below what these models support.

## Pricing

Each model lists input and output price **per million tokens**.

See [Pricing](/models/pricing).