AI · alternatives · 2026

24 alternatives to Gemma 3 inference in pure C++ with Metal acceleration

Pure C++23 LLM inference for Apple Silicon chips. Contribute to ybubnov/metalchat development by creating an account on GitHub.

Below are 24 products that do a similar job, ranked by how close each is in meaning and then by launch-day votes.

1
We made our own inference engine for Apple Silicon▲186
We wrote our inference engine on Rust, it is faster than llama cpp in all of the use cases. Your feedback is very welcomed. Written from scratch with idea that you can add support of any kernel and platform.
2025 · github.com · its alternatives →
2
Tiny-vLLM – high performance LLM inference engine in C++ and CUDA▲205
Build your own high performance LLM inference engine in C++ and CUDA - a smaller version of vLLM - jmaczan/tiny-vllm
May 2026 · github.com · its alternatives →
3
Speeding up LLM inference 2x times (possibly)▲419
Here's a project I've been working on for the last few months. It's a new (I think) algorithm, that allows to adjust smoothly - and in real time - how many calculations you'd like to do during inference of an LLM model. It seems that it's possible to do just 20-25% of weight multiplications instead of all of them, and still get good inference results. I implemented it to run on M1/M2/M3 GPU. The mmul approximation itself can be pushed to run 2x fast before the quality of output collapses. The inference speed is just a bit faster than Llama.cpp's, because the rest of implementation…
2024 · asciinema.org · its alternatives →
4
Open-source engine running Gemma 4 26B in 2 GB RAM on any M-series Mac▲919
Hi HN, I built a specialized inference engine for running 4-bit Gemma 4 26B-A4B-IT on any M-series Mac using about 2 GB of RAM. It is called TurboFieldfare and is written in Swift and Metal. I have always adored on-device AI. It feels like magic that you can run a powerful NN on your Mac or iPhone. So I wanted to push the limits a bit and run a model whose weights don’t fit in memory. The model’s 4-bit quantized weights occupy roughly 14 GB, which makes running it with conventional inference tools almost impossible on an 8 GB or even 16 GB Mac once the OS, applications, and KV cache are…
Jul 2026 · github.com · its alternatives →
5
KVSplit – Run 2-3x longer contexts on Apple Silicon▲272
I discovered that in LLM inference, keys and values in the KV cache have very different quantization sensitivities. Keys need higher precision than values to maintain quality. I patched llama.cpp to enable different bit-widths for keys vs. values on Apple Silicon. The results are surprising: - K8V4 (8-bit keys, 4-bit values): 59% memory reduction with only 0.86% perplexity loss - K4V8 (4-bit keys, 8-bit values): 59% memory reduction but 6.06% perplexity loss - The configurations use the same number of bits, but K8V4 is 7× better for quality This means you can run LLMs with 2-3× longer…
2025 · github.com · its alternatives →
6
Port of OpenAI's Whisper model in C/C++▲399
Hi HN, OpenAI recently released a model for automatic speech recognition called Whisper [0]. I decided to reimplement the inference of the model from scratch using C/C++. To achieve this I implemented a minimalistic tensor library in C and ported the high-level architecture of the model in C++. The entire code is less than 8000 lines of code and is contained in just 2 source files without any third-party dependencies. The Github project is here: https://github.com/ggerganov/whisper.cpp With this implementation I can very easily build and run the model - “make…
2022 · github.com · its alternatives →
7
Gemma 4 Multimodal Fine-Tuner for Apple Silicon▲235
About six months ago, I started working on a project to fine-tune Whisper locally on my M2 Ultra Mac Studio with a limited compute budget. I got into it. The problem I had at the time was I had 15,000 hours of audio data in Google Cloud Storage, and there was no way I could fit all the audio onto my local machine, so I built a system to stream data from my GCS to my machine during training. Gemma 3n came out, so I added that. Kinda went nuts, tbh. Then I put it on the shelf. When Gemma 4 came out a few days ago, I dusted it off, cleaned it up, broke out the Gemma part from the Whisper…
Apr 2026 · github.com · its alternatives →
8
Duplicate 3 layers in a 24B LLM, logical deduction .22→.76. No training▲265
I replicated David Ng's RYS method (https://dnhkng.github.io/posts/rys/) on consumer AMD GPUs (RX 7900 XT + RX 6950 XT) and found something I didn't expect. Transformers appear to have discrete "reasoning circuits" — contiguous blocks of 3-4 layers that act as indivisible cognitive units. Duplicate the right block and the model runs its reasoning pipeline twice. No weights change. No training. The model just thinks longer. The results on standard benchmarks (lm-evaluation-harness, n=50): Devstral-24B, layers 12-14 duplicated once: - BBH Logical Deduction: 0.22 → 0.76…
Mar 2026 · github.com · its alternatives →
9
An educational Local Qwen3 LLM Inference project written in Rust▲15
An educational Rust project for exporting and running inference on Qwen3 LLM family - reinterpretcat/qwen3-rs
2025 · github.com · its alternatives →
10
Revibing nanochat's inference model in C++ with ggml▲5
Recently I wanted to see if I could vibe some serious C++ code. The result is a C++ re-implementation of Andrej Karpathy's nanochat's inferencing part (https://github.com/karpathy/nanochat), built on top of ggml. Unlike llama.cpp, this isn't a standalone binary; it is a C++ library & Python wrapper designed to swap out some core classes within the nanochat pipeline. For playability, I’ve tried to keep the dependencies to a minimum: just ggml, nanobind, and gtest for unit tests. Features and limitations: - A drop-in replacement of nanochat’s `GPT` and `KVCache` classes. So…
Jan 2026 · github.com · its alternatives →
11
Open source machine learning inference accelerators on FPGA▲57
Tom from Tensil here - happy to answer questions! We developed Tensil to bring custom ML accelerators to people who don't have the resources of companies like Google, Facebook and Tesla. Currently, we're focused on supporting convolutional neural network inference on edge FPGA (field programmable gate array) platforms, but we aim to support all model architectures on a wide variety of fabrics for both training and inference. Tensil is different from other ML accelerators in that it is open source and really easy to use. For example, you can generate a custom accelerator with one command: $…
2022 · tensil.ai · its alternatives →
12
GitHub▲7
Bodhisearch/llama.cpp: llm inference in c/c++
2025 · github.com · its alternatives →
13
LLM, a Rust Crate/CLI for CPU Inference of LLMs (LLaMA, GPT-NeoX, etc.)▲45
G'day, HN! I'm one of the maintainers of `llm`. I've been working alongside a trusty group of contributors to bring this project to life, and we're now at a point where we're ready to share it with the world. Large language models (LLMs) are taking the computing world by storm due to their emergent abilities that allow them to perform a wide variety of tasks, including translation, summarization, code generation, and even some degree of reasoning. However, the ecosystem around LLMs is still in its infancy, and it can be difficult to get started with these models. `llm` is a one-stop shop for…
2023 · github.com · its alternatives →
14
I built a tensor library from scratch in C++/CUDA▲119
Hi HN, Over the past few months, I've been building `dsc`, a tensor library from scratch in C++/CUDA. My main focus has been on getting the basics right, prioritizing a clean API, simplicity, and clear observability for running small LLMs locally. The key features are: - C++ core with CUDA support written from scratch. - A familiar, PyTorch-like Python API. - Runs real models: it's complete enough to load a model like Qwen from HuggingFace and run inference on both CUDA and CPU with a single line change[1]. - Simple, built-in observability for both Python and C++. Next on the roadmap is…
2025 · github.com · its alternatives →
15IB
I built a toy TPU that can do inference and training on the XOR problem▲134
We wanted to do something very challenging to prove to ourselves that we can do anything we put our mind to. The reasoning for why we chose to build a toy TPU specifically is fairly simple: - Building a chip for ML workloads seemed cool - There was no well-documented open source repo for an ML accelerator that performed both inference and training None of us have real professional experience in hardware design, which, in a way, made the TPU even more appealing since we weren't able to estimate exactly how difficult it would be. As we worked on the initial stages of this project, we…
2025 · tinytpu.com · its alternatives →
16
SelfHostLLM▲134
A simple calculator that estimates how many concurrent requests your GPU can handle for a given LLM, with shareable results.
2025 · selfhostllm.org · its alternatives →
17
Lightweight Llama3 Inference Engine – CUDA C▲12
Hey, recently I took inspiration from llama.cpp, ollama, and many other similar tools that enable inference of LLMs locally, and I just finished building a Llama inference engine for the 8B model in CUDA C. I recently wanted to explore my newly founded interest in CUDA programming and my passion for machine learning. This project only makes use of the native CUDA runtime api and cuda_fp16. The inference takes place in fp16, so it requires around 17-18GB of VRAM (~16GB for model params and some more for intermediary caches). It doesn’t use cuBLAS or any similar libraries since I wanted to be…
2025 · github.com · its alternatives →
18
LLM Verified with Monte Carlo Tree Search▲102
This is a weekend hack that I'd like to further develop as it's working surprisingly well. Using MCTS, we can explore a space of possible verified programs with an LLM. We check the partial programs at each step, and so steer towards programs that pass the verifier. https://github.com/namin/llm-verified-with-monte-carlo-tree-...
2023 · github.com · its alternatives →
19AT
A tiny LLM running at 21,000 tok/s on a $250 FPGA (Live Demo)▲79
A 3.16M-parameter INT4 transformer running entirely in the on-chip memory of a Xilinx Kria KV260. Zero DRAM in the token loop, 59,965 tok/s on the fabric, bit-exact. Chat with it live.
Aug 2026 · mikeayles.com · its alternatives →
20
Nibble▲102
An attempt at a single pass LLVM frontend in ~3000 lines of C without external dependencies, malloc, or an AST. Included are some graphical examples. The IR isn't perfect, and the README touches on one particular downfall
May 2026 · github.com · its alternatives →
21
Determinstic LLM inference for lowest price Gemma 4, with Windows XP▲9
Open-weight models, fully deterministic. The same answer every time, byte for byte, at the floor price.
Sep 2026 · tokendelivery.ai · its alternatives →
22
Pip install inference, open source computer vision deployment▲71
Deploying vision models is time consuming and tedious. Setting up dependencies. Fixing conflicts. Configuring TRT acceleration. Flashing (and re-flashing) NVIDIA Jetsons. A streamlined, developer-friendly solution for inference is needed. We, the Roboflow team, have been hard at work open sourcing Inference, an open source vision deployment solution. Our solution is designed with developers in mind, offering a HTTP-based interface. Run models on your hardware without having to write architecture-specific inference code. Here's a demo showing how to go from a model to GPU inference on a video…
2023 · github.com · its alternatives →
23
C++ library for reading MacBook lid angle sensor data▲60
C++ library for reading MacBook lid angle sensor data - ufoym/mac-angle
2025 · github.com · its alternatives →
24
Qwen3.6-35B-A3B on a 16 GB M1 Pro with SSD-streamed MoE▲24
Metal-first MoE inference for Apple Silicon with bounded SSD expert streaming and local OpenAI Chat, Responses, and Anthropic Messages endpoints. - hebrus-labs/hebrus
Jul 2026 · github.com · its alternatives →

Also compare

Ranked by how close each launch is in meaning, then by votes. Prices were read from each product’s own site when checked and can change. Refine with your own description →