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 #13462] Feature request: Enable image input for vision-capable cloud models #34643

New Issue
Closed
opened 2026-04-22 18:22:59 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-22 18:22:59 -05:00
Owner
Copy Link

Originally created by @eitelnick on GitHub (Dec 14, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/13462

Cloud models that have vision-capable variants upstream (for example families like Ministral 3 and GLM-4.6/GLM-4.6V) only expose text in Ollama Cloud today (ministral-3:cloud, glm-4.6:cloud).

I’d like to request that, whenever a cloud model has a vision-capable variant, Ollama Cloud exposes image input for it across:

Official UIs (image upload / drag-and-drop)
Native API / SDKs (images: [...] or equivalent)
OpenAI-compatible /v1/chat/completions endpoint
Current vs expected behavior
Current:

Local vision models can accept images via CLI and native API.
Cloud models like ministral-3:cloud and glm-4.6:cloud only support text, even though their families have strong multimodal variants elsewhere.
Expected:

If a cloud model is vision-capable (or has a vision sibling), it should:
Show an image upload option in the UI.
Accept images via images (native API/SDKs).
Accept images via the OpenAI-compatible /v1 API in a documented format.
Why this would help
Parity: Cloud models don’t feel artificially limited vs local vision models.
DX: One simple rule of thumb: “if it’s a vision model (or family), I can always send images, local or cloud.”
Use cases: Code-from-screenshot, UI analysis, diagram understanding, and other visual workflows benefit most from large cloud models.
If there are constraints or a roadmap item for cloud vision support, it’d be great to know. Otherwise, please consider making image/vision support a first-class, consistent feature for all applicable cloud models.

Originally created by @eitelnick on GitHub (Dec 14, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/13462 Cloud models that have vision-capable variants upstream (for example families like Ministral 3 and GLM-4.6/GLM-4.6V) only expose text in Ollama Cloud today (ministral-3:cloud, glm-4.6:cloud). I’d like to request that, whenever a cloud model has a vision-capable variant, Ollama Cloud exposes image input for it across: Official UIs (image upload / drag-and-drop) Native API / SDKs (images: [...] or equivalent) OpenAI-compatible /v1/chat/completions endpoint Current vs expected behavior Current: Local vision models can accept images via CLI and native API. Cloud models like ministral-3:cloud and glm-4.6:cloud only support text, even though their families have strong multimodal variants elsewhere. Expected: If a cloud model is vision-capable (or has a vision sibling), it should: Show an image upload option in the UI. Accept images via images (native API/SDKs). Accept images via the OpenAI-compatible /v1 API in a documented format. Why this would help Parity: Cloud models don’t feel artificially limited vs local vision models. DX: One simple rule of thumb: “if it’s a vision model (or family), I can always send images, local or cloud.” Use cases: Code-from-screenshot, UI analysis, diagram understanding, and other visual workflows benefit most from large cloud models. If there are constraints or a roadmap item for cloud vision support, it’d be great to know. Otherwise, please consider making image/vision support a first-class, consistent feature for all applicable cloud models.
GiteaMirror added the feature request label 2026-04-22 18:22:59 -05:00
GiteaMirror commented 2026-04-22 18:23:03 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Dec 14, 2025):

glm-4.6 is not a vision model. glm-4.6v is a vision model but is not supported in Cloud at the moment.

The following Cloud models are vision capable and accept images via the CLI and Ollama/OpenAI APIs, with the exception of the gemma3 models which respond with an internal server error on the OpenAI API:

qwen3-vl:235b-instruct
qwen3-vl:235b
ministral-3:3b
ministral-3:8b
ministral-3:14b
mistral-large-3:675b
devstral-small-2:24b
gemini-3-pro-preview
gemma3:4b
gemma3:12b
gemma3:27b

I assume they will also accept image uploads through the UI but I don't have access to a Mac/Windows machine to check.

<!-- gh-comment-id:3650049214 --> @rick-github commented on GitHub (Dec 14, 2025): glm-4.6 is not a vision model. glm-4.6v is a vision model but is not supported in Cloud at the moment. The following Cloud models are vision capable and accept images via the CLI and Ollama/OpenAI APIs, with the exception of the gemma3 models which respond with an [internal server error](https://github.com/ollama/ollama/issues/13464) on the OpenAI API: qwen3-vl:235b-instruct qwen3-vl:235b ministral-3:3b ministral-3:8b ministral-3:14b mistral-large-3:675b devstral-small-2:24b gemini-3-pro-preview gemma3:4b gemma3:12b gemma3:27b I assume they will also accept image uploads through the UI but I don't have access to a Mac/Windows machine to check.
GiteaMirror commented 2026-04-22 18:23:03 -05:00
Author
Owner
Copy Link

@eitelnick commented on GitHub (Dec 14, 2025):

Thanks for the response! The documentation, for example for Ministral-3, says "text" only for Cloud models such as ministral-3:14b-cloud. In the UI I can seem to accept image uploads...might just be a documentation fix.

<!-- gh-comment-id:3650231731 --> @eitelnick commented on GitHub (Dec 14, 2025): Thanks for the response! The documentation, for example for [Ministral-3](https://ollama.com/library/ministral-3), says "text" only for Cloud models such as ministral-3:14b-cloud. In the UI I can seem to accept image uploads...might just be a documentation fix.
GiteaMirror commented 2026-04-22 18:23:04 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Dec 14, 2025):

#13468

<!-- gh-comment-id:3650289468 --> @rick-github commented on GitHub (Dec 14, 2025): #13468
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 feature request
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#34643
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?