Footer

Your go-to destination for discovering trending open-source projects and uncovering the insights that matter

Legal

  • Terms and conditions
  • Privacy policy

More

  • Pricing
  • Advertise
  • onedump
  • gosnakego
  • Contact us

Social

  • X (Twitter)
  • GitHub
Trendshift owl
What they build speaks louder than words
© 2026 Trendshift, created by Julian Li with ❤️
SignalAdvertise
Trendshift
SignalAdvertise
Sign in
Trendshift owl

Discover trends that matter

SignalAdvertise
DailyWeeklyMonthlyYearly
Live mentionsTopics
RepositoriesDevelopers
InsightsStats

Omnigraph
Omnigraph
Busbar
Busbar
OfinTech Desk
OfinTech Desk
It's a plan
It's a plan
Bifrost
Bifrost

ewhacc/tandem.cpp

llama.cpp patch: faster MoE offload for models that don't fit in VRAM

#LLM inference
Visit GitHub
Python
99
16
1 contributors
Custom license
FeaturedAdvertise to 180K+ monthly visitors
Omnigraph

Versioned graph database on object storage. Git-style workflows and multimodal retrieval

Busbar

Control what AI can do before it acts

OfinTech Desk

An open-source AI workspace for finance and accounting teams, in which every figure is traced to its source

It's a plan

Open-source Linear alternative

An issue tracker where AI agents take the work you assign them

Bifrost

Fastest Enterprise AI Gateway for Routing, Governing, and Securing all your AI Traffic

Momentum

How fast ewhacc/tandem.cpp is gaining stars, and where it lands on Trendshift and GitHub Trending

Activity and trending

Stars, forks and merged pull requests each day or month, lined up with the repository's rank on Trendshift Trending

Nov 1 – Nov 19, 2026

Stars +94 in 19 days

051015
14

Forks +16 in 19 days

024
4

Merged PRs 0 in 19 days

01

Trendshift Trendingwhere the repository ranked each day it made the top 25

Not in the Trendshift top 25 in this period

Nov 1Nov 4Nov 7Nov 10Nov 13Nov 16Nov 19
The left and right arrow keys step from day to day; Home and End go to the first and the last

Stars, forks, merged pull requests and Trendshift Trending rank, by day
DayStars gainedForks gainedMerged PRs
Sun, Nov 1, 2026+0+00
Mon, Nov 2, 2026+10+20
Tue, Nov 3, 2026+8+10
Wed, Nov 4, 2026+4+10
Thu, Nov 5, 2026+2+00
Fri, Nov 6, 2026+3+00
Sat, Nov 7, 2026+12+40
Sun, Nov 8, 2026+5+20
Mon, Nov 9, 2026+1+00
Tue, Nov 10, 2026+2+00
Wed, Nov 11, 2026+0+00
Thu, Nov 12, 2026+0+00
Fri, Nov 13, 2026+2+00
Sat, Nov 14, 2026+7+10
Sun, Nov 15, 2026+14+10
Mon, Nov 16, 2026+2+20
Tue, Nov 17, 2026+5+00
Wed, Nov 18, 2026+12+20
Thu, Nov 19, 2026+5+00
Total+94+160
  • 1#1 on the list
  • 7Rank that day
  • Not trending

Social mentions

Recent discussions about this repository across the web
  1. Nov 161 post
    • tandem.cpp v0.3.0 gets Qwen3.8-Flash-Next's first token on a 16K prompt in ~15s on a 16GB RTX 5070 Ti, vs 82s in stock @ggml_org.

      Read full on X, posted by @kaimi_local

      @kaimi_local · x.com

  2. 14 days earlier
  3. Nov 23 posts
    • This is the kind of local AI progress I want more of: big open models on an ordinary gaming PC. tandem.cpp patches llama.cpp to keep a MoE model's busiest experts on the GPU. On the author's own RTX…

      Read full on X, posted by @SelfHostedSI

      @SelfHostedSI · x.com

    • 显存装不下的 MoE,让 CPU 分担一部分专家,Qwen3.8-Flash-Next 在一张 4090 上从 29.8 tok/s 跑到了 97.3 tok/s。 这是 tandem.cpp,一组给 llama.cpp 打的补丁,刚开源。测试机是 i5-12400F、64GB DDR4、RTX 4090,模型是 78 GiB 的 IQ3_S 量化,开着…

      Read full on X, posted by @wei_wang

      @wei_wang · x.com

    • 🚀 普通 llama.cpp 跑 Qwen3.8-Flash-Next,4090 上也能到 97 tok/s 了。 新补丁 tandem.cpp:热专家留 GPU,冷的交给 CPU。IQ3_S 约 30→97,还是每个专家都算满。 温馨提示:目前 Linux+CUDA。 来源:GitHub ewhacc/tandem.cpp👇

      Read full on X, posted by @QwenAIFans

      @QwenAIFans · x.com

You might also like

Repositories close to ewhacc/tandem.cpp
FeaturedAdvertise to 180K+ monthly visitors
Busbar

Control what AI can do before it acts

Bifrost

Fastest Enterprise AI Gateway for Routing, Governing, and Securing all your AI Traffic

It's a plan

Open-source Linear alternative

An issue tracker where AI agents take the work you assign them

Omnigraph

Versioned graph database on object storage. Git-style workflows and multimodal retrieval

OfinTech Desk

An open-source AI workspace for finance and accounting teams, in which every figure is traced to its source

GenerelSchwerz/llama.cpp

C++

LLM inference in C/C++

14933

miltos22/llama-wackMall

C++

Caching experts made model agnostic for llama. Drastically speeds up generation speed on MoE models that fully fit in the RAM but partially in Vram. WIP: Same speedup but for models that don't fit in the ram either.

3110

yalun753/moe-l2

Cuda

MoE expert offload for low-VRAM GPUs — run 100B+ MoE models (DeepSeek, Qwen, Mixtral) on 8 GB cards. Expert cache with LRU hot experts, OpenAI-compatible proxy, GGUF multi-shard.

1476

spiritbuun/buun-llama-cpp

C++

Experimental llama.cpp fork for inference research and development

910108

thecodacus/llama.cpp

C++

LLM inference in C/C++

369118

IbrahimSiddiqui007/moe-cache

Python

Run Mixture-of-Experts LLMs (Qwen3, gpt-oss, GLM, DeepSeek) bigger than your RAM with llama.cpp: hot experts cached in RAM, the rest streamed from SSD. Windows (Vulkan/AMD, NVIDIA) and Linux, long context, coding agents.

221

MohammadHaishemKhawaja/hashyy

Python

A full, unpruned 177B MoE model at 11.5 tok/s on one 12 GB GPU. Expert streaming for llama.cpp, with the measurement harness and every dead end.

256

Helldez/BigMoeOnEdge

C++

Run MoE models bigger than your RAM. Frontier-size MoE on a phone, laptop, edge. Lossless, on stock llama.cpp. An adaptive streaming optimizer. Any hardware, any OS, any model, any quantization, behind a local OpenAI-compatible server.

1k118

wici-ai/inferssd

C++

Run models larger than your RAM or VRAM by streaming MoE experts from SSD. Built on llama.cpp.

1.2k108