A high-performance utility to measure and compare Time to First Token (TTFT) latency across major LLM providers in parallel.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install llm-speedtest
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install llm-speedtest using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is LLM Speedtest?
LLM Speedtest is a specialized diagnostic tool built to provide developers with immediate visibility into the performance of various AI model providers. By executing parallel requests, it eliminates the guesswork involved in selecting the most responsive model for a given task. This utility is particularly valuable for developers using Openclaw Skills who require high-speed interactions and need to monitor if specific providers are experiencing regional slowdowns or outages.
The tool focuses on Time to First Token (TTFT), the most critical metric for interactive AI applications. By sending minimal requests to OpenAI, Anthropic, Google, MiniMax, and xAI, it delivers a clear, color-coded leaderboard of current provider speeds without incurring significant API costs.
LLM Speedtest Use Cases
- Comparing real-world provider response times before starting a large coding task.
- Checking if a specific AI model is currently slow or experiencing a partial outage.
- Optimizing automation workflows by selecting the provider with the lowest current latency.
- Benchmarking various Openclaw Skills to ensure optimal environment performance.
How LLM Speedtest Works
- The user initiates the speed test by typing the
/pingcommand within the agent interface. - The skill invokes a shell script that retrieves configured API keys from the local secret manager or environment variables.
- The script fires simultaneous, parallel
curlrequests to all supported LLM endpoints using a minimal prompt. - Latency is measured as the total round-trip time per provider for a single token response.
- The results are aggregated, sorted from fastest to slowest, and rendered with visual badges for quick interpretation.
LLM Speedtest Setup
To get started with this skill, ensure you have the necessary API keys for the providers you want to test. By default, the script sources keys from the pass secret manager, but it can be easily adapted.
# Example: Setting up environment variables if not using pass
export OPENAI_API_KEY="your_key_here"
export ANTHROPIC_API_KEY="your_key_here"
export GEMINI_API_KEY="your_key_here"
Once keys are configured, you can verify the installation by running the ping command in your Openclaw Skills enabled environment.
LLM Speedtest Data Schema & Taxonomy
The skill produces a structured output based on latency thresholds. It organizes results into the following classification schema:
| Icon | Latency Range | Status |
|---|---|---|
| 🟢 | < 2s | Fast |
| 🟡 | 2–5s | Normal |
| 🔴 | 5–30s | Slow |
| ⚫ | 30s+ | Timeout |
Each entry include the provider name, the specific model tested (e.g., Claude 3.5 Sonnet, GPT-4o-mini), and the precise millisecond measurement.
LLM Speedtest Advanced Features
- Parallel execution ensures that the entire benchmark suite finishes in the time it takes for the slowest provider to respond.
- Support for a wide range of top-tier models including Grok 3 Mini, Gemini 2.5 Flash, and MiniMax-M1.
- Cost-efficient design using approximately $0.0001 in credits per full run across all providers.
- Flexible secret management integration allows users to adapt the tool to various Openclaw Skills security setups.
SKILL.md
Loading
Related Openclaw Skills
Claude Watchdog
A robust monitoring tool for the Claude API that delivers real-time status and latency alerts directly to Telegram.
QQ Email Skill
A comprehensive integration for sending and receiving emails through QQ Mail using SMTP and IMAP protocols.
Autonomous Tasks
A self-driven AI worker that autonomously reads goals, generates tasks, executes them, and logs progress.
Patent Disclosure Writer
An automated assistant designed to help users draft professional Chinese patent disclosure and technical documents through a structured interactive dialogue process.
Typefully Social Media Manager
A powerful CLI-driven tool for managing Typefully drafts and scheduling across multiple social platforms.
Typefully Social Media Manager
A powerful automation bridge for Typefully to manage social media drafts and multi-platform scheduling directly via the CLI.
Featured*
