github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 08:30:05 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #10139] OpenAI API: Models with slashes not retrievable #32413

New Issue
Closed
opened 2026-04-22 13:38:25 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-22 13:38:25 -05:00
Owner
Copy Link

Originally created by @SplittyDev on GitHub (Apr 5, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10139

What is the issue?

In the /v1/models/{model} endpoint, models with a / in their name (which is fairly common when downloading from HuggingFace or other sources) are not retrievable individually.

As expected, the model shows up in the model list:

GET /v1/models:

{
    "object": "list",
    "data": [
        {
            "id": "aaa/llava:13b",
            "object": "model",
            "created": 1743869353,
            "owned_by": "aaa"
        },
        {
            "id": "llava:13b",
            "object": "model",
            "created": 1743863049,
            "owned_by": "library"
        }
    ]
}

But when trying to retrieve the model individually, the API always returns 404.

GET /v1/models/aaa/llava:13b (the naive approach):

404 page not found

GET /v1/models/aaa%2Fllava:13b (the proper url-encoded approach):

404 page not found

GET /v1/models/aaa%2Fllava%3A13b (the just-to-be-safe unnecessarily-url-encoded approach):

404 page not found

Relevant links to OpenAI API spec:

Relevant log output

There's nothing in the logs.

OS

macOS

GPU

Apple

CPU

Apple

Ollama version

0.6.4

Originally created by @SplittyDev on GitHub (Apr 5, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10139 ### What is the issue? In the `/v1/models/{model}` endpoint, models with a `/` in their name (which is fairly common when downloading from HuggingFace or other sources) are not retrievable individually. As expected, the model shows up in the model list: `GET /v1/models`: ```json { "object": "list", "data": [ { "id": "aaa/llava:13b", "object": "model", "created": 1743869353, "owned_by": "aaa" }, { "id": "llava:13b", "object": "model", "created": 1743863049, "owned_by": "library" } ] } ``` But when trying to retrieve the model individually, the API always returns `404`. `GET /v1/models/aaa/llava:13b` (the naive approach): ```plain 404 page not found ``` `GET /v1/models/aaa%2Fllava:13b` (the proper url-encoded approach): ```plain 404 page not found ``` `GET /v1/models/aaa%2Fllava%3A13b` (the just-to-be-safe unnecessarily-url-encoded approach): ```plain 404 page not found ``` Relevant links to OpenAI API spec: - `GET /v1/models` (https://platform.openai.com/docs/api-reference/models/list) - `GET /v1/models/{model}` (https://platform.openai.com/docs/api-reference/models/retrieve) ### Relevant log output ```shell There's nothing in the logs. ``` ### OS macOS ### GPU Apple ### CPU Apple ### Ollama version 0.6.4
GiteaMirror added the bug label 2026-04-22 13:38:25 -05:00
GiteaMirror commented 2026-04-22 13:38:29 -05:00
Author
Owner
Copy Link

@chigkim commented on GitHub (Apr 5, 2025):

What are you trying to do? If you want to chat with a model, you have to use chat completion API.
/v1/models/aaa/llava:13b is not OpenAI API.

<!-- gh-comment-id:2781115544 --> @chigkim commented on GitHub (Apr 5, 2025): What are you trying to do? If you want to chat with a model, you have to use chat completion API. `/v1/models/aaa/llava:13b` is not OpenAI API.
GiteaMirror commented 2026-04-22 13:38:30 -05:00
Author
Owner
Copy Link

@SplittyDev commented on GitHub (Apr 6, 2025):

@chigkim It most certainly is part of the OpenAI API: https://platform.openai.com/docs/api-reference/models/retrieve

I'm not trying to use chat completion. OpenAI defines two endpoints for model listing and retrieval:

The issue is that OpenAI models don't have slashes in them. GET /v1/models/gpt-4o works perfectly in the OpenAI API, and so does GET /v1/models/llava:13b in the Ollama OpenAI-compatible API.

But let's take this scenario:

  1. You pull a model from HuggingFace: ollama pull hf.co/TheDrummer/Moistral-11B-v3-GGUF
  2. You call the GET /v1/models endpoint: 
    {
    "object": "list",
    "data": [
        {
            "id": "hf.co/TheDrummer/Moistral-11B-v3-GGUF:latest",
            "object": "model",
            "created": 1743863049,
            "owned_by": "library"
        }
    ]
}
  3. You call the GET /v1/models/hf.co/TheDrummer/Moistral-11B-v3-GGUF:latest endpoint: 
    404 page not found

At first I thought it's because slashes have to be escaped, but turns out that doesn't help. Properly URL-escaped, the model name would be hf.co%2FTheDrummer%2FMoistral-11B-v3-GGUF%3Alatest. But calling the GET /v1/models/hf.co%2FTheDrummer%2FMoistral-11B-v3-GGUF%3Alatest endpoint also results in 404 page not found.

So, it's not actually possible to retrieve a model with slashes in its name in an OpenAI-compatible way, because the URL isn't being decoded correctly, and ollama doesn't find the model id, resulting in a http 404 status error.

<!-- gh-comment-id:2781220627 --> @SplittyDev commented on GitHub (Apr 6, 2025): @chigkim It most certainly _is_ part of the OpenAI API: https://platform.openai.com/docs/api-reference/models/retrieve I'm not trying to use chat completion. OpenAI defines two endpoints for model listing and retrieval: - `GET /v1/models` (https://platform.openai.com/docs/api-reference/models/list) - `GET /v1/models/{model}` (https://platform.openai.com/docs/api-reference/models/retrieve) The issue is that OpenAI models don't have slashes in them. `GET /v1/models/gpt-4o` works perfectly in the OpenAI API, and so does `GET /v1/models/llava:13b` in the Ollama OpenAI-compatible API. But let's take this scenario: 1. You pull a model from HuggingFace: `ollama pull hf.co/TheDrummer/Moistral-11B-v3-GGUF` 2. You call the `GET /v1/models` endpoint: ```json { "object": "list", "data": [ { "id": "hf.co/TheDrummer/Moistral-11B-v3-GGUF:latest", "object": "model", "created": 1743863049, "owned_by": "library" } ] } ``` 3. You call the `GET /v1/models/hf.co/TheDrummer/Moistral-11B-v3-GGUF:latest` endpoint: ``` 404 page not found ``` At first I thought it's because slashes have to be escaped, but turns out that doesn't help. Properly URL-escaped, the model name would be `hf.co%2FTheDrummer%2FMoistral-11B-v3-GGUF%3Alatest`. But calling the `GET /v1/models/hf.co%2FTheDrummer%2FMoistral-11B-v3-GGUF%3Alatest` endpoint also results in `404 page not found`. So, it's not actually possible to retrieve a model with slashes in its name in an OpenAI-compatible way, because the URL isn't being decoded correctly, and ollama doesn't find the model id, resulting in a http 404 status error.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#32413
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?