Open Models · Local Deploy

全球开源模型频道

策展 + Ollama Library 全量 + HF GGUF 热门 合并目录:参数/显存、许可、 Hugging Face / GGUF / Ollama / ModelScope 下载,并跳转 本地算力计算器 。HF 全库十万+——本频道做可跑清单、深链与算力账,不镜像权重。 同步于 2026-08-03T07:33:50Z

合并总量325
策展精选42
同步条数303
本地优先318
含 Ollama238
含 HF/GGUF192

全库入口(全球)

要「全部开源」请用下列权威源检索;本站策展负责选型 + 部署路径 + 算力账

按家族

我的显卡一键筛:

匹配 18 · 展示 18 · 点「设为 A/B」后底部一键对比 · 本站不托管权重

cogito

Community · ~7B · unknown · 本地优先 · ollama-library · 2.1M pulls

Q4 ~4GB舒适 8GB

Cogito v1 Preview is a family of hybrid reasoning models by Deep Cogito that outperform the best available open models of the same size, including counterparts from LLaMA, DeepSeek, and Qwen across most standard benchmarks.

推理Agent对话
ollama run cogito
deepscaler

Community · 1.5b · unknown · 本地优先 · ollama-library · 1.3M pulls

Q4 ~4GB舒适 8GB

A fine-tuned version of Deepseek-R1-Distilled-Qwen-1.5B that surpasses the performance of OpenAI’s o1-preview with just 1.5B parameters on popular math evaluations.

推理数学对话
ollama run deepscaler
deepseek-coder

DeepSeek · ~7B · unknown · 本地优先 · ollama-library · 4.5M pulls

Q4 ~4GB舒适 8GB

DeepSeek Coder is a capable coding model trained on two trillion code and natural language tokens.

代码对话
ollama run deepseek-coder
openthinker

Community · ~7B · unknown · 本地优先 · ollama-library · 1.1M pulls

Q4 ~4GB舒适 8GB

A fully open-source family of reasoning models built using a dataset derived by distilling DeepSeek-R1.

推理对话
ollama run openthinker
deepseek-ocr

DeepSeek · ~7B · unknown · 本地优先 · ollama-library · 503K pulls

Q4 ~6GB舒适 12GB

DeepSeek-OCR is a vision-language model that can perform token-efficient OCR.

视觉OCR
ollama run deepseek-ocr
deepseek-v2.5

DeepSeek · ~7B · unknown · 本地优先 · ollama-library · 283K pulls

Q4 ~4GB舒适 8GB

An upgraded version of DeekSeek-V2 that integrates the general and coding abilities of both DeepSeek-V2-Chat and DeepSeek-Coder-V2-Instruct.

代码对话
ollama run deepseek-v2.5
deepseek-v3.1

DeepSeek · ~7B · unknown · 本地优先 · ollama-library · 718K pulls

Q4 ~4GB舒适 8GB

DeepSeek-V3.1-Terminus is a hybrid model that supports both thinking mode and non-thinking mode.

推理Agent对话
ollama run deepseek-v3.1
deepseek-v4-flash

DeepSeek · 284b/13b · unknown · 本地优先 · ollama-library · 304K pulls

Q4 ~8GB舒适 16GB

DeepSeek-V4-Flash is a preview of the DeepSeek-V4 series, a Mixture-of-Experts model with 284B total parameters and 13B activated, built for efficient reasoning across a 1M-token context window.

推理Agent对话
ollama run deepseek-v4-flash
deepseek-v4-pro

DeepSeek · ~7B · unknown · 本地优先 · ollama-library · 305K pulls

Q4 ~4GB舒适 8GB

DeepSeek-V4-Pro is a frontier Mixture-of-Experts model with a large context window and three reasoning modes.

推理Agent对话
ollama run deepseek-v4-pro
r1-1776

Community · ~7B · unknown · 本地优先 · ollama-library · 414K pulls

Q4 ~4GB舒适 8GB

A version of the DeepSeek-R1 model that has been post trained to provide unbiased, accurate, and factual information by Perplexity.

推理对话
ollama run r1-1776

与竞品对齐的能力

  • ApX / VRAM 计算器类:本站 /tools/local-deploy 覆盖显存适配、多卡 TP、产出与 vs API 费用。
  • Ollama Library / HF Hub:每条模型直链下载与 ollama run,不二次托管权重。
  • Model Picker(按 GPU 筛模型):列表「显存上限」筛选 + 计算器预填。
  • 价格/中转对照:有 OpenRouter id 的可回 /models · /official-api
开源模型部署频道 · HF / Ollama / GGUF · GrokCode 倍率榜