# ikawrakow/ik_llama.cpp

> A llama.cpp fork with additional SOTA quantization types and improved CPU/GPU performance for running LLMs locally. It supports new models and inference techniques that landed here before upstream.

- Magnitude: 1.2 out of 10 — Steady
- Stars: 3,271 total · +2 stars today, ≈ 4 by evening
- Star trust: star growth looks organic
- Category: Language models · Language: C++ · License: MIT · Created: 2024-06-27 · Last push: 2026-09-29
- GitHub: https://github.com/ikawrakow/ik_llama.cpp · Page: https://gitnova.dev/en/r/ikawrakow/ik_llama.cpp

## Useful for

- Run a local llama-server with a GGUF model on CPU or CUDA
- Build the binary with AVX-512 flags to enable IQK quantized GEMM kernels
- Deploy inference in Docker using a prebuilt ghcr.io image

## Why it’s here

- 2 stars so far today, about 4 expected by the end of the day.
- GitHub Trending C++ today: #15, +6 stars.

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 470
- Issues and pull requests: 2,343
- Watchers: 35
- Average over the last week: 3 per day
- Usual pace: 5 per day
- Stars in the last hour (measured): 0
- Latest release: t0002 (2025-07-22)

## Stars per day, last 30 days (oldest → newest, today is partial)

2026-08-31 … 2026-09-29: 12, 3, 5, 4, 4, 3, 5, 4, 12, 3, 2, 3, 7, 1, 5, 6, 2, 8, 4, 5, 2, 2, 5, 0, 1, 4, 0, 5, 6, 2

## Spotted in now

- GitHub Trending C++ today: #15, +6 stars

## More in this category

1. **firelex/jeff** — 8.2 · Early signal · Language models · Python · +255 stars today, ≈ 486 by evening
   Small fine-tuned Qwen3.5 and Gemma 4 models for zero-shot classification: given a situation description and a list of options, they return a calibrated probability for each option in a single forward pass.
   Full card: https://gitnova.dev/en/r/firelex/jeff.md
2. **VectifyAI/PageIndex** — 7.4 · Breakout · Language models · Python · +285 stars today, ≈ 538 by evening
   A vectorless RAG engine that builds a hierarchical tree index of a document and retrieves by LLM reasoning over that tree, the way a human navigates a report. Aimed at long professional PDFs such as financial, legal and technical documents.
   Full card: https://gitnova.dev/en/r/VectifyAI/PageIndex.md
3. **Niko1221/Strata** — 7.2 · Early signal · Language models · C++ · +183 stars today, ≈ 387 by evening
   A C++ local inference engine that runs the 125B MoE model Qwen3.8-Flash-Next on a regular PC with one NVIDIA GPU (12-24 GB) and 64 GB RAM, exposing an OpenAI/Anthropic-compatible API on localhost.
   Full card: https://gitnova.dev/en/r/Niko1221/Strata.md
4. **PostHog/jeeves** — 6.6 · Early signal · Language models · Python · +83 stars today, ≈ 119 by evening
   Jeeves is a reasoning model built on Qwen3.5-9B (LoRA + pointer head), trained with SFT and CISPO, that thinks before answering and returns calibrated probabilities for yes/no, multiple-choice and rating questions via a Jev-compatible API.
   Full card: https://gitnova.dev/en/r/PostHog/jeeves.md
5. **ollaya-dev/ollaya** — 5.6 · Early signal · Language models · Rust · +40 stars today, ≈ 90 by evening
   A local runtime for decision models: pulls and serves classification and routing models behind a TypeSafe-compatible API. Like Ollama, but for models that return probabilities instead of text.
   Full card: https://gitnova.dev/en/r/ollaya-dev/ollaya.md

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-09-29 13:28 UTC, updated every 30 minutes.