# speculative-decoding

> Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques.

Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques. Use when optimizing inference speed (1.5-3.6× speedup), reducing latency for real-time applications, or deploying models with limited compute. Covers draft models, tree-based attention, Jacobi iteration, parallel token generation, and production deployment strategies.

- Type: Skill
- Made by: Orchestra-Research
- Category: AI & agents
- Formats: Skills
- Version: e556e4a
- Source: https://github.com/orchestra-research/ai-research-skills/tree/HEAD/19-emerging-techniques/speculative-decoding
- Page: https://pluginboard.ai/skills/speculative-decoding

## Works with

- ChatGPT
- Claude
- Gemini
- GitHub Copilot
- Cursor
- Claude Code
- Codex
- Antigravity
- Gemini CLI
- Replit
- OpenClaw
- OpenCode
- Hermes Agent

## Install

### ChatGPT

In ChatGPT: Download the speculative-decoding folder and zip it, then Plugins → Skills → Create → Upload from your computer. Business and Enterprise plans.

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

### Claude

In Claude (web or desktop): Download the speculative-decoding folder and zip the folder itself, then upload it in Customize → Skills.

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

### Gemini

In Gemini: Download the speculative-decoding folder, then at gemini.google.com click Customize → Skills → Create → Upload and pick the folder (or a .zip of it).

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

### GitHub Copilot

Run in your terminal:

```shell
npx -y degit Orchestra-Research/AI-Research-SKILLs/19-emerging-techniques/speculative-decoding ~/.copilot/skills/speculative-decoding
```

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

### Cursor

Run in your terminal:

```shell
npx -y degit Orchestra-Research/AI-Research-SKILLs/19-emerging-techniques/speculative-decoding ~/.cursor/skills/speculative-decoding
```

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

### Claude Code

Run in your terminal:

```shell
npx -y degit Orchestra-Research/AI-Research-SKILLs/19-emerging-techniques/speculative-decoding ~/.claude/skills/speculative-decoding
```

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

Or all of AI Research Skills (Orchestra), in Claude Code:

```text
/plugin marketplace add Orchestra-Research/AI-Research-SKILLs
/plugin install rag@ai-research-skills
```

### Codex

Run in your terminal:

```shell
npx -y degit Orchestra-Research/AI-Research-SKILLs/19-emerging-techniques/speculative-decoding ~/.agents/skills/speculative-decoding
```

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

### Antigravity

In Antigravity: Copy the speculative-decoding folder into ~/.gemini/config/skills (Antigravity app and IDE) or ~/.gemini/antigravity-cli/skills (Antigravity CLI).

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

### Gemini CLI

Run in your terminal:

```shell
npx -y degit Orchestra-Research/AI-Research-SKILLs/19-emerging-techniques/speculative-decoding ~/.gemini/skills/speculative-decoding
```

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

### Replit

Run in your terminal:

```shell
npx -y degit Orchestra-Research/AI-Research-SKILLs/19-emerging-techniques/speculative-decoding .agents/skills/speculative-decoding
```

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

### OpenClaw

Run in your terminal:

```shell
npx -y degit Orchestra-Research/AI-Research-SKILLs/19-emerging-techniques/speculative-decoding ~/.openclaw/skills/speculative-decoding
```

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

### OpenCode

Run in your terminal:

```shell
npx -y degit Orchestra-Research/AI-Research-SKILLs/19-emerging-techniques/speculative-decoding ~/.config/opencode/skills/speculative-decoding
```

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

### Hermes Agent

Run in your terminal:

```shell
npx -y degit Orchestra-Research/AI-Research-SKILLs/19-emerging-techniques/speculative-decoding ~/.hermes/skills/speculative-decoding
```

View the folder on GitHub: https://github.com/Orchestra-Research/AI-Research-SKILLs/tree/HEAD/19-emerging-techniques/speculative-decoding

---
From Pluginboard (https://pluginboard.ai).