v0.38.0

Mesh-LLM/mesh-llmv0.38.0Mar 20, 2026by michaelneale

AI Summary

This release adds vision and multimodal support to the mesh, allowing models to process images alongside text. It introduces three new vision-native models and updates the default 'Mini' model, while retiring a legacy API application.

Key Highlights

Automatic download and launch of multimodal projectors for vision models
Three new vision-native models (Qwen3.5-0.8B, 4B, 9B)
UI updates allowing image attachment for vision-capable models
Retirement of the legacy `mesh-llm-api` app in favor of the console

Breaking Changes

Retirement of the Fly API app (`mesh-llm-api`)

New Features

Vision support via automatic `--mmproj` flag
New vision models: Qwen3.5-0.8B, 4B, 9B, and Qwen3.5-27B
UI image attachment button for vision models
Vision badge in model dropdown
Updated `/v1/models` endpoint to return vision capabilities
Updated `/api/status` endpoint to include vision field

Full Release Notes

## Vision/Multimodal Support

Models with vision capability now automatically get their multimodal projector (mmproj) downloaded and launched with `--mmproj`.

### New
- **Vision support**: Catalog models can declare an mmproj file. Download fetches it alongside the model. llama-server launches with `--mmproj` automatically.
- **3 new vision models**: Qwen3.5-0.8B, 4B, 9B (all vision-native)
- **Qwen3.5-27B** tagged as vision-capable with mmproj
- **UI image attach**: Button appears for vision-capable models. Attach images, see previews, send with messages.
- **👁 badge**: Vision models marked in model dropdown and dashboard
- **`/v1/models` capabilities**: Returns `["text", "vision"]` for vision models
- **`/api/status` vision field**: `vision: true/false` per model

### Changed
- Mini swapped from Qwen3-8B to Qwen3.5-9B (same size, gains vision)
- Fly API app (`mesh-llm-api`) retired — console serves both UI and API

### Mesh
- 3 hosts: Mini (Qwen3.5-9B 👁), Studio (MiniMax-M2.5), James (Qwen2.5-72B)
- Console: https://www.mesh-llm.com/chat