Poolside33B (3B active)Q4_K_M 下约 21 GB 显存

Laguna XS 2.1 33B-A3B — 显存、速度与本地部署

作者: Jakub Rusinowski · 最后更新:

33B total parameters with only 3B activated per token, which is what lets a model scoring 70.9% on SWE-Bench Verified sit on a single 24GB card. Mixed sliding-window + global attention (3:1 across 40 layers), sigmoid per-head gating, FP8 KV cache, 256K context. Released under OpenMDW-1.1 — the Linux Foundation / NVIDIA permissive model-weights framework — not Apache 2.0, so check the terms if you are redistributing.

Laguna XS 2.1 33B-A3B 在 Q4_K_M 下约需 21 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。

逻辑82
创意62
编程94

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。

量化显存速度(估算)适配
Q2_K
2.63 bpw
11.6 GB~210 tok/s可运行
Q3_K_M
3.41 bpw
14.9 GB~194 tok/s可运行
Q4_K_M
4.83 bpw
20.7 GB~171 tok/s可运行
Q5_K_M
5.67 bpw
24.2 GB~24 tok/s需卸载
Q6_K
6.56 bpw
27.9 GB~22 tok/s需卸载
Q8_0
8.50 bpw
35.9 GB~20 tok/s需卸载
F16
16.00 bpw
66.8 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. Laguna XS 2.1 33B-A3B 显存计算器 →

运行 Laguna XS 2.1 33B-A3B

目录中能运行 Laguna XS 2.1 33B-A3B 的最便宜 GPU 是 AMD Radeon RX 7900 XTX (24 GB).

购买此硬件 AMD Radeon RX 7900 XTX 24GB — 24 GB VRAM · 355 W board power立即云端部署 RunPod 上的 RTX 4090 — 低至 $0.34/小时 · 价格核实于 2026-07

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。

联盟营销声明: 本页部分链接为联盟推广链接——如果你通过它们购买,LLM Configurator 可能会获得佣金,而你无需支付任何额外费用。作为亚马逊联盟成员(Amazon Associate),LLM Configurator 会从符合条件的购买中获得收益。
AMD Radeon RX 7900 XTX 24GB
24 GB VRAM · 355 W board power
2026年价格波动较大——请以当前商品页价格为准。

如何运行 Laguna XS 2.1 33B-A3B

安装 Ollama,然后运行:

ollama run poolside-laguna
Hugging Face 上的权重: poolside/Laguna-XS-2.1 ↗

规格

Preview — The model is released, but these specs are thin or rest on a single source. Individual fields may be wrong.

Preview — The model is released, but these specs are thin or rest on a single source. Individual fields may be wrong.

参数量
33 Billion (3B active)
上下文窗口
262,144
架构
Mixture-of-Experts (sliding-window + global attention, FP8 KV cache)
提供商
Poolside
许可证
OpenMDW-1.1
Practical deployment
Single consumer GPU
规格量化
Q4_K_M
系统内存
32 GB
记录更新于
2026-08-15
许可证OpenMDW-1.1允许商业使用

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。

最适合agentic codingsoftware engineeringrepo levelconsumer gpulong documents
基准分数来源
SWE-Bench Verified70.9 / 100 %已发布 · https://poolside.ai/blog/introducing-laguna-xs-2-1
SWE-bench Multilingual63.1 / 100 %已发布 · https://poolside.ai/blog/introducing-laguna-xs-2-1

我的 GPU 能运行 Laguna XS 2.1 33B-A3B 吗?

Laguna XS 2.1 33B-A3B — 常见问题

How much VRAM does Laguna XS 2.1 33B-A3B need?

About 21 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 Laguna XS 2.1 33B-A3B run on an RTX 4090 (24 GB)?

Yes. Laguna XS 2.1 33B-A3B needs about 21 GB at Q4_K_M, inside a 24 GB card, at an estimated 171 tokens/sec.

How do I run Laguna XS 2.1 33B-A3B locally?

Install Ollama and run `ollama run poolside-laguna`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.