ollama - 💡(How to fix) Fix [Question] Why is GLM-5 only available as a cloud model and not for local download? [5 comments, 5 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
ollama/ollama#14724Fetched 2026-04-08 00:32:30
View on GitHub
Comments
5
Participants
5
Timeline
6
Reactions
0
Author
Timeline (top)
commented ×5labeled ×1
RAW_BUFFERClick to expand / collapse

Hi Ollama team,

I am a regular user of Ollama, and I have a technical question regarding how certain models are listed in your library.

While comparing different open-source models, I noticed a difference in availability. For instance, on the qwen3.5 page, users have the choice between locally downloadable versions (e.g., 27b, 35b, 122b with their respective weights in GB) and a cloud version (qwen3.5:cloud) (Also, as a side question, what size version do you use for the cloud version most of the time? The largest one? For example, 122b in my case, or is there an even larger one that's not accessible?).

However, on the glm-5 page (https://ollama.com/library/glm-5), only the glm-5:cloud version is offered.

Since GLM-5 is also an open-source model, I would love to understand why it isn't available for local download on Ollama.

Out of curiosity, what are the technical criteria or the decision-making process for determining which models can be executed locally versus those that are limited to your cloud integration?

Thank you very much for your time and the fantastic work you do!

Thank you in advance, best regards

extent analysis

Fix Plan

To address the issue of missing local downloadable versions for certain models, we need to update the model configuration and add support for local downloads.

Steps:

  • Update the model metadata to include local downloadable versions
  • Add a download endpoint for the model weights
  • Update the frontend to display the local downloadable versions

Example Code:

# models.py
class Model:
    def __init__(self, name, cloud_version, local_versions):
        self.name = name
        self.cloud_version = cloud_version
        self.local_versions = local_versions

# Add local downloadable versions for GLM-5
glm_5 = Model(
    name="glm-5",
    cloud_version="glm-5:cloud",
    local_versions=[
        {"version": "27b", "size": "10GB"},
        {"version": "35b", "size": "15GB"},
        {"version": "122b", "size": "30GB"}
    ]
)

# downloads.py
from flask import send_file

@app.route("/models/<model_name>/download/<version>")
def download_model(model_name, version):
    model = get_model(model_name)
    for local_version in model.local_versions:
        if local_version["version"] == version:
            return send_file(local_version["weight_path"], as_attachment=True)
    return "Model not found", 404

# frontend.js
fetch("/models/glm-5")
    .then(response => response.json())
    .then(data => {
        const localVersions = data.local_versions;
        localVersions.forEach(version => {
            const downloadLink = document.createElement("a");
            downloadLink.href = `/models/glm-5/download/${version.version}`;
            downloadLink.textContent = `Download ${version.version}`;
            document.body.appendChild(downloadLink);
        });
    });

Verification

To verify that the fix worked, check the model page for the presence of local downloadable versions. Try downloading a version to ensure that the download endpoint is working correctly.

Extra Tips

  • Make sure to update the model weights and metadata regularly to ensure that the local downloadable versions are up-to-date.
  • Consider adding a caching mechanism to reduce the load on the download endpoint.
  • Use a secure protocol (e.g., HTTPS) to serve the model weights and metadata.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING