ik_llama.cpp
Fewer than 5 people recommend it so far, so there's no written verdict. Here is what commenters said.
3recommend
0warn against
Jan 6, 2026first mention
Aug 23, 2026latest mention
Why people recommend it
“Hello, it's the internet calling, today is that day.”
“ikllama.cpp is a fork of llama.cpp which specializes on CPU inference”
Show 1 more
“Have you tried anything with https://codeberg.org/ikawrakow/illama”
Caveats and dissent
No one in these comments warned against it.
Often compared with
Threads where it came up
- Why your local LLM feels dumber than it is 2 mentions
- Laguna S 2.1 1 mention
- Running Gemma 4 26B at 5 tokens/SEC on a 13-year-old Xeon with no GPU 1 mention
- Qwen 3.6 27B is the sweet spot for local development 1 mention
- Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding? 1 mention
- A 10 year old Xeon is all you need (for 26B-A4B MTP Drafters without GPU) 1 mention
Data: ik-llama-cpp.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 12
“Q8_0 from llama.cpp / ik_llama.cpp should also work better than FP8 from vllm”
“(and "ik_llama.cpp" in very few cases)”
“It works, thanks to https://github.com/ikawrakow/ik_llama.cpp/pull/1911”
“The fix is open upstream as PR #2138”
“Hello, it's the internet calling, today is that day.”
“switching to ik_llama.cpp fork”
“ik_llama includes llama-sweep-bench”
“backends (llama.cpp, ik_llama.cpp) have been updated”
“did tests on the quality of turboquant-3 compared to IQ4 quantized models”
“I went with ik_llama.cpp”
“ikllama.cpp is a fork of llama.cpp which specializes on CPU inference”
“Have you tried anything with https://codeberg.org/ikawrakow/illama”