ollama - 💡(How to fix) Fix Why is Video Input STILL not supported? (Qwen 3.5 / Gemma 4 are already here, yet PRs are ignored)

It is frankly baffling that we are still asking this question.

The open-source LLM landscape has rapidly evolved. Models like Qwen 3.5, Gemma 4, and MiniCPM already have robust, native support for video inputs. Yet, Ollama—which is supposed to be the go-to tool for running local models—has become the primary bottleneck preventing users from utilizing these state-of-the-art multimodal capabilities.

What is even more frustrating than the lack of the feature is the repository management regarding this issue:

1. Ignoring Community Efforts: PRs that actually attempt to solve this (like #12962) are left to rot in limbo without meaningful feedback or merging.
2. Deceptive Issue Management: Issue #8419 was explicitly asking for video support, yet it was arbitrarily closed as "completed" by maintainers without a single line of code actually enabling video input. Closing an issue doesn't make the demand disappear.

The community is doing the work. The models already support it. Why is the Ollama team actively ignoring essential feature PRs while falsely marking feature requests as "completed"?

We don't need vague "the backend is undergoing changes" responses that stretch for months. We need to know:

• Is there a concrete timeline for video support?
• Why are working PRs for this exact feature being ignored?

Please stop closing legitimate feature requests to keep the issue tracker clean, and start leveraging the community's PRs so we can actually use the models to their full potential.