EXAONE 4.0 1.2B — 显存、速度与本地部署
作者： Jakub Rusinowski · 最后更新：
The on-device EXAONE 4.0: 1.3B parameters with the same reasoning and direct-answer modes, and a 65,536-token context. LG's official Q4_K_M GGUF is 0.8 GB. Research and education use only.
EXAONE 4.0 1.2B 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。在 Apple Silicon 上，这部分来自统一内存。
按量化级别的显存与速度
计算基准：NVIDIA RTX 4090 (24 GB)。包含 8K 上下文的 KV 缓存，因此数值高于上方的主数字。
| 量化 | 显存 | 显存 | 速度（估算） | 适配 |
|---|---|---|---|---|
| Q2_K 2.63 bpw | 1.7 GB | ~319 tok/s | 可运行 | |
| Q3_K_M 3.41 bpw | 1.8 GB | ~303 tok/s | 可运行 | |
| Q4_K_M 4.83 bpw | 2.1 GB | ~278 tok/s | 可运行 | |
| Q5_K_M 5.67 bpw | 2.2 GB | ~265 tok/s | 可运行 | |
| Q6_K 6.56 bpw | 2.4 GB | ~252 tok/s | 可运行 | |
| Q8_0 8.50 bpw | 2.7 GB | ~228 tok/s | 可运行 | |
| F16 16.00 bpw | 3.9 GB | ~168 tok/s | 可运行 |
黑色标记 = NVIDIA RTX 4090 (24 GB) 上的可用显存。 估算来自内存带宽屋顶线模型，详见 方法说明页. EXAONE 4.0 1.2B 显存计算器 →
运行 EXAONE 4.0 1.2B
目录中能运行 EXAONE 4.0 1.2B 的最便宜 GPU 是 Intel Arc B570 (10 GB).
或在 Vast.ai 比较，低至 $0.35/小时 (typical low · varies)
作为亚马逊联盟成员，我们从符合条件的购买中获得收入。云 GPU 链接为推荐链接——我们可能获得佣金，您无需额外付费。
如何运行 EXAONE 4.0 1.2B
安装 Ollama，然后运行：
ollama run exaone-4规格
Verified — Checked against the primary source — the model card or the vendor spec page — and corroborated by a second independent source.
- 参数量
- 1.2 Billion
- 上下文窗口
- 65,536
- 架构
- Dense transformer (GQA)
- 提供商
- LG AI Research
- 许可证
- EXAONE AI Model License 1.2 - NC
- Availability
- Weights downloadable
- Practical deployment
- Edge device (phone, SBC)
- 规格量化
- Q4_K_M
- 系统内存
- 4 GB
- 记录更新于
- 2026-11-11
Research / non-commercial only — this licence does NOT permit shipping a commercial product.
另请参阅 EXAONE 4.0 32B
质量与使用场景
评分由模型作者或独立评测方发布——衡量质量而非吞吐量，并非我们实测。
EXAONE 4 的其他尺寸
EXAONE 4.0 1.2B — 常见问题
How much VRAM does EXAONE 4.0 1.2B need?
About 2 GB at Q4_K_M — quantized weights plus framework overhead, before any KV cache. The cache grows with context length and is added on top; the table above folds it in. Apple Silicon counts unified memory toward the same figure.
Does EXAONE 4.0 1.2B run on an RTX 4090 (24 GB)?
Yes. EXAONE 4.0 1.2B needs about 2 GB at Q4_K_M, inside a 24 GB card, at an estimated 278 tokens/sec.
How do I run EXAONE 4.0 1.2B locally?
Install Ollama and run `ollama run exaone-4`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.
What other sizes does EXAONE 4 come in?
EXAONE 4.5 33B (21 GB), EXAONE 4.0 32B (20 GB), EXAONE 4.0 1.2B (2 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.