[Feature]: Add /v1/models API endpoint to list models #532
Description
Prerequisites
- I have searched existing issues and discussions to avoid duplicates
Problem to solve
When /v1/models or /models is queried from clients - bifrost does not return the list of models.
LLM API providers use the /v1/models or /models endpoint to provide clients a list of models. This is often used to populate model drop down lists in client tools.
curl https://bifrost.your.domain/v1/models
<!DOCTYPE html><!--build--><html lang="en"><head><meta charSet="utf-8"/><meta name="viewport" content="width=device-width, initial-scale=1"/><link rel="preload" href="/_next/static/media/4cf2300e9c8272f7-s.p.woff2" as="font" crossorigin="" type="font/woff2"/><link rel="preload" href="/_next/static/media/93f479601ee12b01-s.p.woff2" as="font" crossorigin="" type="font/woff2"/><link rel="stylesheet" href="/_next/static/css/59786e30c3453a61.css" data-precedence="next"/><link rel="preload" as="script" fetchPriority="low" href="/_next/static/chunks/webpack-07c7d5bd191e9c71.js"/><script src="/_next/static/chunks/4bd1b696-c023c6e3521b1417.js" async=""></script><script src="/_next/static/chunks/255-2ea93159d9bee7fd.js" async=""></script><script src="/_next/static/chunks/main-app-2ecc3bf42ae2d122.js" async=""></script><script src="/_next/static/chunks/82-9cd62844ea761b4e.js" async=""></script><script src="/_next/static/chunks/194-cf0d37d90654fe28.js" async=""></script><script src="/_next/static/chunks/691-c47bd2fca38cb22d.js" async=""></script><script src="/_next/static/chunks/520-b89056c58a467df4.js" async=""></script><script src="/_next/static/chunks/619-ba102abea3e3d0e4.js" async=""></script><script src="/_next/static/chunks/834-aad6f4e4087827a9.js" async=""></script><script src="/_next/static/chunks/52-9e5c42ee942dd760.js" async=""></script><script src="/_next/static/chunks/522-71a58c8ec82d9bd6.js" async=""></script><script src="/_next/static/chunks/105-dea6bda13558f5ca.js" async=""></script><script src="/_next/static/chunks/app/layout-6bbdae7071e6833f.js" async=""></script><script src="/_next/static/chunks/app/page-16d38047884ba52a.js" async=""></script><meta name="next-size-adjust" content=""/><link rel="icon" href="/favicon.ico" type="image/x-icon" sizes="32x32"/><script src="/_next/static/chunks/polyfills-42372ed130431b0a.js" noModule=""></script></head><body class="__variable_188709 __variable_9a8899 antialiased"><div hidden=""><!--$--><!--/$--></div><style>
:root {
--bprogress-color: #188410;
--bprogress-height: 4px;
--bprogress-spinner-size: 18px;
--bprogress-spinner-animation-duration: 400ms;
--bprogress-spinner-border-size: 2px;
--bprogress-box-shadow: 0 0 10px #188410, 0 0 5px #188410;
--bprogress-z-index: 99999;
--bprogress-spinner-top: 15px;
--bprogress-spinner-bottom: auto;
--bprogress-spinner-right: 15px;
--bprogress-spinner-left: auto;
}
...
This causes client applications to fail to list models, e.g:
Proposed solution
Add support for the /v1/models standard endpoint for listing models.
For example when doing a curl https://your-server.your.domain/v1/models (with llama-swap, vllm etc...) you should get back something like:
{
"data": [
{
"created": 1759206400,
"id": "Kimi-Dev-72B-UD-Q3_K_XL-49k",
"object": "model",
"owned_by": "llama-swap"
},
{
"created": 1759206400,
"id": "Mistral-Small-3.2-24B-Instruct-2506-UD-Q5_K_XL-32k",
"object": "model",
"owned_by": "llama-swap"
},
{
"created": 1759206400,
"id": "Mistral-Small-3.2-24B-Instruct-2506-UD-Q5_K_XL-64k",
"object": "model",
"owned_by": "llama-swap"
},
{
"created": 1759206400,
"id": "deepseek-v3.1-terminus-ud-tq1_0-32k",
"object": "model",
"owned_by": "llama-swap"
},
{
"created": 1759206400,
"id": "embeddinggemma-300m-q8_0-2k",
"object": "model",
"owned_by": "llama-swap"
},
{
"created": 1759206400,
"id": "fine-tuned-transcripts-llama3.2-3b-q4km-64k",
"object": "model",
"owned_by": "llama-swap"
}
]
}Another example (Ollama):
{
"object": "list",
"data": [
{
"id": "embeddinggemma:latest",
"object": "model",
"created": 1757628380,
"owned_by": "library"
},
{
"id": "hf.co/SaisExperiments/granite-4.0-tiny-preview-Q8_0-GGUF:latest",
"object": "model",
"created": 1752230934,
"owned_by": "SaisExperiments"
},
{
"id": "hf.co/unsloth/SmolLM3-3B-128K-GGUF:UD-Q6_K_XL",
"object": "model",
"created": 1752203646,
"owned_by": "unsloth"
},
{
"id": "hf.co/unsloth/Devstral-Small-2507-GGUF:UD-Q6_K_XL",
"object": "model",
"created": 1752190437,
"owned_by": "unsloth"
}
]
}Alternatives considered
The closest existing issue I could find is #493 but /v1/models (or simply /models) is not a management endpoint.
Area(s)
Transports (HTTP)
Additional context
No response
Metadata
Metadata
Assignees
Labels
Type
Projects
Status