Home › Use-case lists › Local Large Model Running Tools
🖥️ Local Large Model Running Tools (5)
Run large models on your own computer or server so data never leaves the machine and there is no network dependency. This list filters tools and models related to local running, model quantisation and local inference from the hot list, sorted by popularity. Content updates automatically with the hot list.
Data updated Oct 18, 2026 · Prices not verified; check official websites
1. PolyStrata
PolyStrata is a tool for running large language models locally. It supports multiple models including Qwen3, GLM 5.3, DeepSeek V4, LFM2.5 and Gemma 4. It can be installed and run with one click on Windows and Linux, and provides a local API compatible with OpenAI/Anthropic. Key features include: Supports running multiple large language models locally on a PC, including Qwen3.8-Flash-Next, GLM 5.3 Flash, DeepSeek V4 Flash, LFM2.5, Qwen3.6 and Gemma 4; Provides one-click installation and running on Windows and Linux platforms; Built-in local API service compatible with OpenAI and Anthropic. Suitable for developers, researchers and individual users who need to run large language models privately on a local PC, develop applications with integrated AI capabilities, or use a local API.
PolyStrata: full overview and alternatives → · Official website / project page
2. deepseek-local-guide-ru
Complete Russian-language guide detailing how to locally install and run DeepSeek V4 and R1 models on PC, Mac and servers, supporting multiple methods including Ollama, llama.cpp, LM Studio, vLLM, Open WebUI, API, RAG, Agent and fine-tuning. Key features include: Complete Russian guide detailing how to locally install and run the DeepSeek V4 and R1 models on PC, Mac and servers; Supports one-click running of DeepSeek-R1 distilled models via Ollama; Supports running the DeepSeek V4-Flash model using llama.cpp. Suitable for Russian-speaking developers, researchers and enthusiasts who want to run DeepSeek models privately on local PCs, Macs or servers.
deepseek-local-guide-ru: full overview and alternatives → · Official website / project page
3. d3
vllm-sr/d3 is an open-source model repository based on the vLLM inference engine, focused on efficient deployment and inference of large language models. Key features include: Supports text or JSON input and can attach multiple images and videos simultaneously; A single call can handle multiple decision types: multiple-choice, yes/no, scoring, and returns the probability for each answer; Outputs decision probabilities directly without generating text. Developers who need to perform efficient multimodal decision inference, suitable for building selection, judgement or scoring applications.
d3: full overview and alternatives → · Official website / project page
4. Qwen3.8-Flash-Next-FP4-Local
NVIDIA-released Qwen3.8B Flash Next FP4 quantised version, a lightweight large language model for local execution. Key features include: FP4 quantised version of Qwen3.8B Flash Next, supporting local execution; Quantised using Quantisation-Aware Distillation (QAD) technology; Supports vision encoder, hybrid attention, MoE, gated residual streams and n-gram embeddings. Developers and researchers who need to deploy and run lightweight multimodal large language models on local NVIDIA hardware (such as RTX PRO 6000 Blackwell or DGX Spark).
Qwen3.8-Flash-Next-FP4-Local: full overview and alternatives → · Official website / project page
5. Underdog-Saluki-27B-1.0
A 2-bit GGUF quantised model based on Qwen3.8-27B, with a size under 8GB. It focuses on preserving tool-calling capability and can be run directly with llama.cpp. Key features include: A 2-bit GGUF quantised model based on Qwen3.8-27B, size 7.89GB; Focuses on preserving tool-calling capability and supports parallel tool calls; Supports text generation and can achieve image understanding via an optional vision add-on. Users who need to run tool-calling and multimodal tasks on large models but have limited hardware resources. Suitable for developers and researchers deploying llama.cpp locally.
Underdog-Saluki-27B-1.0: full overview and alternatives → · Official website / project page
FAQ
Which tools are recommended in “Local Large Model Running Tools”?
This list includes 5 tools: PolyStrata, deepseek-local-guide-ru, d3, Qwen3.8-Flash-Next-FP4-Local, Underdog-Saluki-27B-1.0. From the AI Hot List (updated every 6 hours): PolyStrata, deepseek-local-guide-ru, d3, Qwen3.8-Flash-Next-FP4-Local, Underdog-Saluki-27B-1.0.
Are these tools free?
Prices and free tiers change often and this site has not verified them one by one. Please check each tool's official website. PolyStrata, deepseek-local-guide-ru, d3, Qwen3.8-Flash-Next-FP4-Local, Underdog-Saluki-27B-1.0 are published openly on GitHub / Hugging Face; see their project pages for licence and usage details.
How often is this list updated?
The curated part is maintained by our editors; the hot-list part updates automatically with the AI Hot List every 6 hours. This data was last updated on Oct 18, 2026.
More lists
🧩 Essential AI tools for product managers🎮 AI tools for indie game development💻 AI Programming / Vibe Coding Tools🎬 AI Video Generation Tools✍️ AI Writing Tools🤖 AI Agent Building Tools🎙️ AI Meeting Notes and Transcription Tools
Newsletter
One email a week
One email a week: the 3 must-see tools from the hot list.
Free. No spam. Unsubscribe in one click.