K2 Horizon MoVA 36B-A4B — 显存、速度与本地部署
作者： Jakub Rusinowski · 最后更新：
A mixture-of-experts model with Mixture-of-Values attention: 37.4B parameters stored (the card rounds to 36B), about 4B active per token, so it generates quickly for its size. It scores 25 on the Artificial Analysis Intelligence Index, second only to Qwen3.8-27B among models that fit one consumer GPU when checked. Its card reports Terminal-Bench 2.1 58.6 against 44.9 for Qwen3.6-35B-A3B. The official Q4_K_M GGUF is 22.4 GB, and its plain grouped-query cache adds about 6.4 GB at 32K context, so this is a 32 GB-card model; a 24 GB card does not hold it at useful context. Apache 2.0, 512K native context.
K2 Horizon MoVA 36B-A4B 在 Q4_K_M 下约需 23 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 | 14.7 GB | ~198 tok/s | 可运行 | |
| Q3_K_M 3.41 bpw | 18.4 GB | ~180 tok/s | 可运行 | |
| Q4_K_M 4.83 bpw | 25 GB | ~23 tok/s | 需卸载 | |
| Q5_K_M 5.67 bpw | 28.9 GB | ~21 tok/s | 需卸载 | |
| Q6_K 6.56 bpw | 33.1 GB | ~20 tok/s | 需卸载 | |
| Q8_0 8.50 bpw | 42.1 GB | ~17 tok/s | 需卸载 | |
| F16 16.00 bpw | 77.2 GB | — | 放不下 |
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. K2 Horizon MoVA 36B-A4B 显存计算器 →
运行 K2 Horizon MoVA 36B-A4B
目录中能运行 K2 Horizon MoVA 36B-A4B 的最便宜 GPU 是 NVIDIA GeForce RTX 5090 (32 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 K2 Horizon MoVA 36B-A4B
Ollama 模型库暂无标签。请用 llama.cpp 运行官方 GGUF：文件声明的架构为 k2-horizon，llama.cpp 主线已支持；较旧的 Ollama 或 LM Studio 版本可能无法加载。模型卡也提供 vLLM 和 SGLang 部署方式。
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
- 37.4 Billion (4B active)
- 上下文窗口
- 524,288
- 架构
- Mixture-of-Experts (100 routed + 1 shared) with Mixture-of-Values attention
- 提供商
- MBZUAI Institute of Foundation Models
- 许可证
- Apache 2.0
- Practical deployment
- Single consumer GPU
- 规格量化
- Q4_K_M
- 系统内存
- 64 GB
- 记录更新于
- 2026-10-08
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。
| 基准 | 分数 | 来源 |
|---|---|---|
| Terminal-Bench 2.1 | 58.6 / 100 % | 已发布 · https://huggingface.co/IFM/K2-Horizon-MoVA-36B-A4B |
K2 Horizon 的其他尺寸
K2 Horizon MoVA 36B-A4B — 常见问题
How much VRAM does K2 Horizon MoVA 36B-A4B need?
About 23 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 K2 Horizon MoVA 36B-A4B run on an RTX 4090 (24 GB)?
Yes. K2 Horizon MoVA 36B-A4B needs about 23 GB at Q4_K_M, inside a 24 GB card, at an estimated 23 tokens/sec.
How do I run K2 Horizon MoVA 36B-A4B locally?
Ollama 模型库暂无标签。请用 llama.cpp 运行官方 GGUF：文件声明的架构为 k2-horizon，llama.cpp 主线已支持；较旧的 Ollama 或 LM Studio 版本可能无法加载。模型卡也提供 vLLM 和 SGLang 部署方式。 Running the published tag would send your prompts to a hosted GPU rather than your own machine.
What other sizes does K2 Horizon come in?
K2 Horizon 3.7B (4 GB), K2 Horizon 7B (6 GB), K2 Horizon MoVA 36B-A4B (23 GB), K2 Horizon 375B-A23B (230 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.