Mostly fixes, but OpenAI-compatible sampler defaults changed and cloud /api/usage reportedly dropped percentage fields. Validate client integrations before upgrading production.
## What happens With the stock `qwen2.5` library template, each tool in the `
### What is the issue? I am building a private local AI assistant called **Maistos Stack Hub**, using Ollama, Open WebUI, and a Docker-based Python execution sandbox on Windows 11...
### What is the issue? On an RTX Spark N1X laptop (HP OmniBook X 14, 128 GB unified memory, 80 GB dedicated), Ollama in WSL2 offloads only part of a model that fits in GPU memory....
I accidentally re-pulled qwen3.5:9b and its tag now points to a newer manifest. Today, ollama show qwen3.5:9b --modelfile / /api/tags reports digest 56671c2ab9385f9cfcb404638e32cd6...
**Note: this is an Ollama Cloud API issue, not a local runtime issue — the OS/GPU/CPU fields below are intentionally left empty.** ## What is the issue? Ollama Cloud (`https://ol...
I think Ollama should gain an API to let applications check if it is up-to-date (by checking the installed version against the latest one by probing the official channels, e.g: git...
Could you please consider reducing the filesize of Windows Installer. The size is so big because of all the bundeled GPU runtimes. I propose to create an installer for each GPU cat...
## Summary `GET https://ollama.com/api/usage` (Bearer API key) no longer returns the session/weekly usage percentages. It now returns only request-count buckets. The percentages s...
If v0.40.14 breaks your environment, revert container image tag to previous stable release: