ollama/ollama v0.34.1
GGUF model creation now requires using llama.cpp tooling for safetensor conversion and quantization.
ProofCode ↗
Key points
- MLX safetensors ollama create no longer experimental.
- Runaway repeat token detection now requires 100 repeat tokens for reduced false positives (e.g. OCR)
- /api/tags is much faster on large model libraries (3.1 s → 294 ms cold in testing), and model capabilities are now reported consistently.
- Deprecated typicalp: it can no longer be set when creating new models, existing GGUF models retain support.
Sources (1)
- [1]ollama/ollama v0.34.1GitHub: ollama/ollama · Sep 14, 10:14 PM
GGUF model creation now requires using llama.cpp tooling for safetensor conversion and quantization.
* MLX safetensors `ollama create` no longer experimental.
Extractive summary: sentences quoted from the sources.
Before this
- Sep 5, 2026ollama/ollama v0.34.0
- Sep 2, 2026ollama/ollama v0.33.3
- Aug 26, 2026ollama/ollama v0.33.1
- Aug 21, 2026ollama/ollama v0.33.0
- Aug 19, 2026ollama/ollama v0.32.15
- Jun 9, 2026Introducing Gemma 4 12B: a unified, encoder-free multimodal model