The model table in docs/self-host.md#engines only mentions Qwen2.5. Run ./scripts/aha-chat.sh with a Llama 3.x and a Mistral instruct GGUF (LLAMACPP_HF_MODEL=...) and add rows with the model, quantization, and result. Note whether agentic mode with wiki_search works.
The model table in
docs/self-host.md#enginesonly mentions Qwen2.5. Run./scripts/aha-chat.shwith a Llama 3.x and a Mistral instruct GGUF (LLAMACPP_HF_MODEL=...) and add rows with the model, quantization, and result. Note whether agentic mode withwiki_searchworks.