AION
Open-source releaseEfficiency & Inference · Large Language Models · MLOps, Tooling & Infrastructure1 source · Sep 23, 2026

ollama/ollama v0.34.4

Qwen 3.8 prompt processing is faster on Apple Silicon.

ChatGPTQwenAppleGemmaCodexllama.cpp
← Back to brief

Key points

  • Structured outputs on thinking models now apply in a single pass, making them faster and more reliable.
  • Fixed intermittent "model not found" errors with a large local library
  • Fixed the macOS app becoming unresponsive when checking if ChatGPT or Codex is running.
  • Gemma 4 on Apple Silicon now picks the best image resolution per image, keeping more detail in high-resolution images.

Sources (1)

  • [1]ollama/ollama v0.34.4
    GitHub: ollama/ollama · Sep 23, 02:24 AM
    - Qwen 3.8 prompt processing is faster on Apple Silicon.
    - Structured outputs on thinking models now apply in a single pass, making them faster and more reliable.

Extractive summary: sentences quoted from the sources.