Pestle-27B-Ternary
把 27B 模型压缩为单个 8.48 GB GGUF 的三值权重本地模型,面向医疗问答、生物医学检索、编程与通用助手
✅ 优点与优势 (Pros)
- • 27B 级模型压缩为单个 8.48 GB GGUF(三值权重)
- • 强医疗基准:MedQA 89.79、MedMCQA 68.85、PubMedQA 76.70
- • 本地用 Mortar 运行于 Apple Silicon / CUDA / CPU
- • 通用能力保留:MMLU-Redux 83.53、GSM8K 93.25、HumanEval+ 87.20
- • 最长 262K 上下文,可选视觉输入
❌ 缺点与限制 (Cons)
- • 仅研究预览,不可用于临床
- • 需自行构建/运行独立 Mortar 运行时
- • 压缩相对全精度 FP16 损失部分准确率
💰 价格方案 (Pricing)
Free (Open Weights, Apache-2.0)
价格详情收集自公开渠道,可能存在变动。请访问官方网站以获取实时最新价格。
最后更新:2026 年 7 月 · 9bests 编辑评测
✅ 谁适合使用 Pestle-27B-Ternary
- • 27B 级模型压缩为单个 8.48 GB GGUF(三值权重)
- • 强医疗基准:MedQA 89.79、MedMCQA 68.85、PubMedQA 76.70
- • 本地用 Mortar 运行于 Apple Silicon / CUDA / CPU
- • 通用能力保留:MMLU-Redux 83.53、GSM8K 93.25、HumanEval+ 87.20
- • 最长 262K 上下文,可选视觉输入
⚠️ 谁可能需要再考虑
- • 仅研究预览,不可用于临床
- • 需自行构建/运行独立 Mortar 运行时
- • 压缩相对全精度 FP16 损失部分准确率
🎯 常见使用场景
文献与论文精读
实验追踪
综合与对齐研究
⚖️ Pestle-27B-Ternary 对比 SkillSpector
| Pestle-27B-Ternary | SkillSpector | |
|---|---|---|
| 评分 | 4.35/5 | 4.3/5 |
| 价格 | Free (Open Weights, Apache-2.0) | Free (Open Source) |
| 核心优势 | 27B 级模型压缩为单个 8.48 GB GGUF(三值权重) | 提升工作流效率 |
❓ 常见问题
Pestle-27B-Ternary 免费吗?
+
Pestle-27B-Ternary 提供免费方案(Free (Open Weights, Apache-2.0))。付费版可解锁更高额度与高级功能。
Pestle-27B-Ternary 最适合做什么?
+
Pestle-27B-Ternary 最适合27B 级模型压缩为单个 8.48 GB GGUF(三值权重)以及强医疗基准:MedQA 89.79、MedMCQA 68.85、PubMedQA 76.70。把 27B 模型压缩为单个 8.48 GB GGUF 的三值权重本地模型,面向医疗问答、生物医学检索、编程与通用助手
Pestle-27B-Ternary 和 SkillSpector 有什么区别?
+
Pestle-27B-Ternary(4.35/5)与 SkillSpector(4.3/5)需求重叠。Pestle-27B-Ternary 在27B 级模型压缩为单个 8.48 GB GGUF(三值权重)上更突出,而 SkillSpector 擅长提升工作流效率。请根据你的优先项选择。
🔄 Pestle-27B-Ternary 的最佳替代方案
相关工具推荐SkillSpector
NVIDIA-developed AI agent skill security scanner that detects vulnerabilities, malicious patterns, and safety risks in AI agent skills.
Memento
Self-hosted agentic search and LLM wiki over your email archive, turning decades of emails into a personal wiki with People, Projects, Concepts, and Newsletters dimensions.
Reyn
Always-on local-first AI that watches your screen, journals your work, and provides instant search across everything you've worked on.
ModelMap
ModelMap (modelmap.tech) is an interactive 3D visualization that turns AI model benchmark scores into explorable shapes. Each model's performance across public benchmarks is rendered as a 'spiky' 3D form — longer spikes mean higher scores — parsed live from Hugging Face model cards. Built on an open-source '3D Graph' library, it offers a flight-simulator-style interface (WASD to fly, mouse to look, click a spike to zoom, hover for tooltips) for browsing model data in space rather than static tables. A hidden Star Wars-themed mini-game underscores its goal of making model analysis more playful. It's a free, browser-based research toy — novel for building intuition, though it has drawn technical criticism on how it represents scores.