EXAONE 4 — LG AI Research 的本地 AI 模型
作者： Jakub Rusinowski · 最后更新：
LG AI Research 的 EXAONE 4 系列：EXAONE 4.0（32B 与 1.2B，2025 年 7 月）提供推理模式和直接回答模式；EXAONE 4.5（2026 年 4 月）是 LG 首个开放的视觉语言模型。支持英语、韩语和西班牙语。许可仅限研究和教育用途；商业使用需另行取得 LG 的许可。
变体
EXAONE 4 最小的变体在 Q4_K_M 下约需 2 GB 显存——量化权重加框架开销，不含 KV 缓存。
| 模型 | Q4 下显存 | 显存 | 上下文 | 运行 |
|---|---|---|---|---|
| EXAONE 4.5 33B → 33B | ~20.8 GB | 262,144 | ollama run exaone-4 | |
| EXAONE 4.0 32B → 32B | ~20.1 GB | 131,072 | ollama run exaone-4 | |
| EXAONE 4.0 1.2B → 1.2B | ~1.6 GB | 65,536 | ollama run exaone-4 |
显存为 Q4_K_M 下的量化权重加开销，与 GPU 与显存检测器使用同一引擎计算。
如何在本地运行 EXAONE 4
安装 Ollama，然后拉取标签。
ollama run exaone-4在上方选择一个尺寸，查看它自己的显存、速度估算和安装命令。
许可证
Research / non-commercial only — this licence does NOT permit shipping a commercial product.
适用于: EXAONE 4.5 33B, EXAONE 4.0 32B, EXAONE 4.0 1.2B推荐 GPU
目录中能在本地运行 EXAONE 4（至少 2 GB 显存）的最便宜 GPU 是 Intel Arc B570 (10 GB).
EXAONE 4 — 常见问题
How much VRAM does EXAONE 4 need?
EXAONE 4 needs about 2 GB VRAM at Q4_K_M quantization for its smallest variant. Variants: EXAONE 4.5 33B (21 GB, Q4_K_M); EXAONE 4.0 32B (20 GB, Q4_K_M); EXAONE 4.0 1.2B (2 GB, Q4_K_M). On Apple Silicon, unified memory counts toward this requirement.
Can I run EXAONE 4 on an RTX 4090 (24 GB)?
Yes — EXAONE 4 runs on an RTX 4090 (24 GB) and other 24 GB cards such as the RTX 3090. Smaller variants also fit comfortably on 8–16 GB GPUs at Q4_K_M.
What quantization should I use for EXAONE 4?
Q4_K_M is the best balance of quality and VRAM for EXAONE 4 in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
How do I run EXAONE 4 with Ollama?
EXAONE 4 has no local Ollama tag — the published tag is cloud-hosted, so running it sends your prompts to a hosted GPU rather than your own machine.