AI models
A plain-language profile of every model you can run yourself: parameter count, hardware it needs, which quantisation fits. No rankings, no prices — just what it takes to run.
Models listed · 8
Save your device profile once and this page will tell you whether it runs on your machine.
Add your deviceNo account? Try a one-off check- 7B
Qwen2.5-7B
阿里云通义千问（Qwen）
拿来微调出自己下游模型的 7B 基座权重。
- Min VRAM
- 6G
- Download size
- —
- Architecture
- Dense
EstimatedContext 128KSave your device profile once and this page will tell you whether it runs on your machine.
- 7B
Qwen2.5-VL 7B
阿里云通义千问（Qwen）
能看图的多模态模型，7B 级，12GB 显存起步。
- Min VRAM
- 12G
- Download size
- —
- Architecture
- Dense
EstimatedContext 128KSave your device profile once and this page will tell you whether it runs on your machine.
- 70B
Llama 3.3 70B Instruct
Meta
需要 48GB 级显存的大家伙：本地能力的天花板之一。
- Min VRAM
- 47G
- Download size
- —
- Architecture
- Dense
EstimatedContext 128KSave your device profile once and this page will tell you whether it runs on your machine.
- 8B
Llama 3.1 8B Instruct
Meta
能在 8GB 显存的普通显卡上本地跑起来的中等模型。
- Min VRAM
- 6G
- Download size
- —
- Architecture
- Dense
EstimatedContext 128KSave your device profile once and this page will tell you whether it runs on your machine.
- 30B-A3B
Qwen3-30B-A3B
阿里云通义千问（Qwen）
总量 30B、每次只激活 3B 的 MoE 模型：大模型的能力，小模型的速度。
- Min VRAM
- 18G
- Download size
- —
- Architecture
- MoE (mixture of experts)
EstimatedContext 128KSave your device profile once and this page will tell you whether it runs on your machine.
- 7B
Qwen2.5 7B Instruct
阿里云通义千问（Qwen）
中文理解明显更好的 7B 级模型，8GB 显存能跑。
- Min VRAM
- 6G
- Download size
- —
- Architecture
- Dense
EstimatedContext 128KSave your device profile once and this page will tell you whether it runs on your machine.
- 7B
DeepSeek-R1-Distill-Qwen-7B
DeepSeek（深度求索）
把"先想再答"的能力压进 7B，本地也能做推理题。
- Min VRAM
- 6G
- Download size
- —
- Architecture
- Dense
EstimatedContext 128KSave your device profile once and this page will tell you whether it runs on your machine.
- 12B
FLUX.1 [schnell]
Black Forest Labs
本地文生图的 12B 模型，8GB 显存可用低显存模式跑。
- Min VRAM
- 8G
- Download size
- —
- Architecture
- Diffusion
EstimatedSave your device profile once and this page will tell you whether it runs on your machine.