RepoAlerts

JakeATX/llamAmpere

App

Can I trust it?

No red flags in our checks.

  • Stars from real accounts No data

    We haven't checked its stars for fakes yet.

  • Active OK

    Last change 2 days ago.

  • Licence OK

    MIT: you can use it for free, also in a commercial product. Keep the licence notice.

Quick checks from public data. They help, but they are no guarantee.

What it is

A fork of llama.cpp tuned for NVIDIA RTX 30-series Ampere GPUs, especially the RTX 3090 and 3090 Ti. It runs Qwen3.8 models, including Qwen3.8-27B with its built-in MTP draft head and a Flash-Next variant, with support for long context and low-memory configurations.

Who it's for

Developers and enthusiasts running local LLMs on consumer RTX 30-series GPUs who need long context or faster inference

How to use it

Install

scripts/flash-next-server.sh -m models/flash-next/qwen3.8-flash-next-Swift1.5-ATX-AI3-00001-of-00002.gguf

From its README. Check it there before running it.

Ready-made installers, when the project publishes them, are on its latest release.

Downloads (if any)

Ask your AI to set it up

Paste this into Claude, ChatGPT or your coding assistant.

Help me install and set up JakeATX/llamAmpere in my project. Follow its README: https://github.com/JakeATX/llamAmpere. Explain each step and check it is safe before running anything.

Star curve

383From 112 to 383 stars between 26 Sept 2026 and 16 Nov 2026.

Alternatives

Similar projects named in its profile. Each link searches the directory.

Get alerts for repos like this

We'll email you when something in Models & inference starts rising. Free, once a week, only what matches.

What arrives: one email on Mondays with the repos that match, never more. See last week's. Privacy

For maintainers

Add this badge to your README. It shows the repo's place on the radar, updates every hour and links back to this page.

[![JakeATX/llamAmpere on RepoAlerts](https://repoalerts.com/badge/JakeATX/llamAmpere.svg)](https://repoalerts.com/repo/JakeATX/llamAmpere?ref=badge)
Share on X

  • freriadoc/Avalon

    A fork of llama.cpp tuned for running the Qwen3.8-Flash-Next 125B MoE model on a single consumer AMD GPU with experts offloaded to system RAM. It uses an expert cache in VRAM, CPU computation of missi

    LibraryModels & inferenceC++ · 436 stars · +50 this week · 38 days old

  • vixhal-baraiya/qwenGPU

    An experimental model: Qwen3.5-2B with a SIMT GPU compiled into its weights as 45 extra layers. The model executes CUDA kernels compiled to PTX, with its Gated DeltaNet layers holding registers and me

    ModelModels & inferencePython · 120 stars · +2 this week · 15 days old

  • Betelgeuse-AI/Betelgeuse

    A llama.cpp fork that runs mixture-of-experts models larger than GPU memory on an RTX 3060 12 GB. Dense weights stay on the GPU, while routed experts are paged from local NVMe into a bounded VRAM slot

    AppModels & inferenceC++ · 154 stars · +31 this week · 16 days old

  • manuelgscom/inside-llm

    A native binary that loads one or two GGUF models via llama.cpp and shows in the browser per-token probabilities, discarded alternatives, sampler effects, and the token where two models' outputs diver

    AppModels & inferenceTypeScript · 62 stars · +62 this week · 1 day old

  • samir-sarwar/samir-vllm

    A learning-oriented LLM inference engine written in C++17 and CUDA that runs Llama 3.2 1B Instruct. It loads safetensors weights to the GPU, implements tokenizer, prefill and decode with paged KV cach

    ResourceModels & inferenceC++ · 223 stars · +188 this week · 81 days old · HN 80 pts

  • divagr18/QwenPU

    A Qwen3.5-4B model modified with a 32-bit RISC-V CPU and a vector unit compiled into its weights, which runs Karpathy's stories260K Llama as a C program inside the model's forward passes. It is a rese

    ModelModels & inferencePython · 139 stars · +8 this week · 16 days old