Olmo 3 7B — 显存、速度与本地部署
作者： Jakub Rusinowski · 最后更新：
The small Olmo 3, in Base, Instruct and Think builds. Runs on an 8 GB card at Q4. Its value is provenance: every token it was trained on is published.
Olmo 3 7B 在 Q4_K_M 下约需 5 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)。仅含权重与开销：该模型架构未公开，因此未计入 KV 缓存。
| 量化 | 显存 | 显存 | 速度（估算） | 适配 |
|---|---|---|---|---|
| Q2_K 2.63 bpw | 3.1 GB | ~166 tok/s | 可运行 | |
| Q3_K_M 3.41 bpw | 3.8 GB | ~144 tok/s | 可运行 | |
| Q4_K_M 4.83 bpw | 5 GB | ~117 tok/s | 可运行 | |
| Q5_K_M 5.67 bpw | 5.8 GB | ~105 tok/s | 可运行 | |
| Q6_K 6.56 bpw | 6.5 GB | ~95 tok/s | 可运行 | |
| Q8_0 8.50 bpw | 8.2 GB | ~78 tok/s | 可运行 | |
| F16 16.00 bpw | 14.8 GB | ~47 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. Olmo 3 7B 显存计算器 →
运行 Olmo 3 7B
目录中能运行 Olmo 3 7B 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 Olmo 3 7B
安装 Ollama，然后运行：
ollama run olmo-3规格
Corroborated — Two or more independent sources agree on these figures, but the model card itself was not retrieved. Treat the numbers as good rather than confirmed. 尚未确认： context.
- 参数量
- 7 Billion
- 上下文窗口
- 65,536
- 架构
- Dense
- 提供商
- Ai2 (Allen Institute for AI)
- 许可证
- Apache 2.0
- Practical deployment
- Edge device (phone, SBC)
- 规格量化
- Q4_K_M
- 系统内存
- 16 GB
- 记录更新于
- 2026-09-19
Commercial use permitted. No usage restrictions beyond attribution.
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。
Olmo 3 的其他尺寸
Olmo 3 7B — 常见问题
How much VRAM does Olmo 3 7B need?
About 5 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 Olmo 3 7B run on an RTX 4090 (24 GB)?
Yes. Olmo 3 7B needs about 5 GB at Q4_K_M, inside a 24 GB card, at an estimated 117 tokens/sec.
How do I run Olmo 3 7B locally?
Install Ollama and run `ollama run olmo-3`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.
What other sizes does Olmo 3 come in?
Olmo 3 7B (5 GB), Olmo 3 32B (20 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.