Model Hierarchy

Cost-optimize AI agent operations by routing tasks to appropriate models based on complexity. Use this skill when: (1) deciding which model to use for a task, (2) spawning sub-agents, (3) considering cost efficiency, (4) the current model feels like overkill for the task. Triggers: "model routing", "cost optimization", "which model", "too expensive", "spawn agent".

346 stars 21 forksAI, LLMs & Agents

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add zscole/model-hierarchy-skill --skill model-hierarchy
Manual install
git clone --depth 1 https://github.com/zscole/model-hierarchy-skill.git /tmp/model-hierarchy-skill
mkdir -p ~/.claude/skills
cp -r /tmp/model-hierarchy-skill/. ~/.claude/skills/model-hierarchy

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Model Hierarchy

Route tasks to the cheapest model that can handle them. Most agent work is routine.

## Core Principle

**80% of agent tasks are janitorial.** File reads, status checks, formatting, simple Q&A. These don't need expensive models. Reserve premium models for problems that actually require deep reasoning.

## Model Tiers

### Tier 1: Cheap ($0.10-0.50/M tokens)

| Model | Input | Output | Best For |
|-------|-------|--------|----------|
| DeepSeek V3 | $0.14 | $0.28 | General routine work |
| GPT-4o-mini | $0.15 | $0.60 | Quick responses |
| Claude Haiku | $0.25 | $1.25 | Fast tool use |
| Gemini Flash | $0.075 | $0.30 | High volume |
| GLM 5 (Zhipu) | (OpenRouter Z.AI) | (OpenRouter Z.AI) | Routine + moderate text; 200K context; **text-only** — do not use for image/vision |
| Kimi K2.5 (Moonshot) | $0.45 | $2.25 | Routine + moderate; 262K context; **multimodal (text + image + video)** |

**Text-only models (e.g. GLM 5):** Do not use for any task that requires image input or vision — no photo analysis, screenshots, image-generation tools, or document/chart vision. Route to a vision-capable model (e.g. Kimi K2.5, GPT-4o, Gemini, Claude with vision, GLM-4.5V/4.6V).

**Vision-capable Tier 1/2 (e.g. Kimi K2.5):** Use for routine or moderate tasks that may involve images — screenshots, photo analysis, docs, image-generation orchestration — without moving to premium vision models.

### Tier 2: Mid ($1-5/M tokens)

| Model | Input | Output | Best For |
Read the full skill on GitHub

Details

Skill name
model-hierarchy
Repository
zscole/model-hierarchy-skill
License
MIT
Repo updated
16 February 2026
Listed in
VoltAgent/awesome-agent-skills

FAQ

What does the Model Hierarchy skill do?

Cost-optimize AI agent operations by routing tasks to appropriate models based on complexity. Use this skill when: (1) deciding which model to use for a task, (2) spawning sub-agents, (3) considering cost efficiency, (4) the current model feels like overkill for the task. Triggers: "model routing", "cost optimization", "which model", "too expensive", "spawn agent".

How do I install the Model Hierarchy skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add zscole/model-hierarchy-skill --skill model-hierarchy. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Model Hierarchy popular?

It lives in zscole/model-hierarchy-skill, which has 346 GitHub stars and 21 forks, and is included in 1 curated skill list.

Related skills

Registry Broker

471

hashgraph-online/registry-broker-skills

Search and chat with AI agents across the Universal Agentic Registry via the Hashgraph Online Registry Broker API. Use when discovering agents, starting conversations, finding incoming messages, or registering new agents.

AI, LLMs & Agents

Stop AI Slop Jp

476

iKora128/stop-ai-slop-jp

AIで書いた日本語を、人間が書いた文章に戻す。下書き、編集、レビューで使う。全角ダッシュや偏愛語だけでなく、主体の不在、命題型H2、壮大化、両論併記、リズムの均一さも直す。

AI, LLMs & Agents17

Superlocalmemory

269

qualixar/superlocalmemory

AI agent memory with mathematical foundations. Store, recall, search, and manage memories locally. Local data root; optional networked features have separate behavior.

AI, LLMs & AgentsIn 1 list37

Serenity Bottleneck Hunter

412

Mrjie7205/serenity-bottleneck-hunter

给定一个投资主题/趋势,复用交易者 Serenity(@aleabitoreddit)的"供应链瓶颈逆向映射"方法论,独立挖出被市场忽视的上游瓶颈股(而非分析他已喊过的标的)。当用户给出一个主题(如"AI 数据中心电力""人形机器人""HBM 内存")并想要候选标的+论证时使用。仅供研究教育,非投资建议。

AI, LLMs & Agents72

AI Short Drama

345

A-cat-with-carrots/OnlyShot

精细化 AI 短剧 IP 创作技能（v0.6.0）。三阶段架构：Phase 1 创作（剧本+ref图，反复迭代）→ Phase 1.5 分镜图（每 grid 1-4 张候选静态图 = 视频首帧，工业级核心层）→ Phase 2 出片（按集解锁，4 模自动选）。v0.6.0 关键升级：① ref 库工艺偏置铁律（现代摩天楼易拟物，古建筑必出 chibi 人体，前期 IP 设计阶段就要避坑）② 即梦 5.0 失败模式 + 敏感词清单（3 类 fail 区分 / prompt 1500 字硬上限 / 反派词替换表 / 暧昧词清单）③ 分镜图 8 段 prompt 模板（CHARACTER/BACKGROUND/ACTION/SCENE/CAMERA/LIGHT/TEXT/STYLE）+ NOT humans 子句必加。沿用 v0.3.0 升级：Phase 1.5 分镜图层、4 模视频、ref 5-8 最优。v0.2.0：bash → Python subprocess、36 grid × 4-10s 变奏、红果必爆 7 招、工业级 ref 库 80-150 张、单 prompt 300-500 字。务必触发：用户提到短剧、微短剧、竖屏剧、AI 短剧、AI 漫剧、剧本创作、分镜、即梦/Seedance 出片、红果/番茄/抖音 IP 改编、爽剧、重生、穿越、赘婿、追妻、神医相师、AI 漫剧奇

AI, LLMs & Agents34

Wallstreet Explainer

419

HeyChengC/wallstreet-explainer

华尔街式的财经/科技解读短片（1–3 分钟、16:9、没有主持人出镜）：用真实的人、真实的新闻和官方画面作证据；当事人的实拍抠像替代主持人（人站在巨字前面、人身后换背景、横推换人）；图形舞台（厚玻璃圆盘、大数字、白底图表）放在一个有前后层的 2.5D 世界里，用镜头把钱和关系演出来，每一镜都接住上一镜往下推。教的是设计思路、动效原理（为什么不像 PPT、不像 AI）和风格语言，外加一套工具（找片下片看片抽帧、新闻原文截图、GPU 逐帧抠人、临时配音和时间线、逐帧渲染、2.5D 相机小工具、镜头运动自检、混音），不给可以照抄的画面模板。用户说"wallstreet-explainer""做一条财经/科技解读""把一笔账讲清楚""巨头砸钱/估值/营收/资金循环这类题目"时使用；有真人口播出镜的片子不用这个。

AI, LLMs & Agents86

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.