Ollama v0.40.16 Released
#Ollama #Apple #OpenAI #NVIDIA
Ollama v0.40.16 has been released. It includes the addition of a forward truncation option for openai API requests, support for CUDA 13.5 and SM 103 for ARM64 on Windows, and memory efficiency improvements for Qwen3.8 Flash Next in MLX.
Ollama v0.40.16 has been released. Regarding openai changes, a forward truncation option for response requests was added, and /v1/models now returns an empty array when no models are installed.
Other updates include a version bump of llama.cpp to b12177. For Windows, CUDA 13.5 and SM 103 for ARM64 were added. In MLX, Qwen3.8 Flash Next was made memory efficient, and ripple preview support was added to the MLX runner. Additionally, Claude Code routing and retired model errors in launch were fixed.
This item is a summary. The full, checkable version is at the original source:
Read the original at Ollama ↗