by gengchaogit · Agent Tool · ★ 97
Last updated: · Indexed by AgentSkillsHub · Auto-synced every 8h
🚀 本地大模型推理速度测试工具 v3 (chao魔改版) 🇬🇧 View in English (查看英文版) | 🇨🇳 查看中文版 ✨ 在线体验最新纯前端版(无需下载/部署,点开即用): 👉 测速页面: https://gengchaogit.github.io/llmspeedtest/ 👉 全网排行榜: https://gengchaogit.github.io/llmspeedtest/LLMSpeedTestv3Leaderboard.html 简介 本项目是一个功能强大的本地大模型推理性能测试工具,旨在帮助用户快速、便捷地测试本地部署的各种大模型推理服务的 Prefill(预填充)和 Decode(解码/输出)性能。 提供两种版本: 纯前端版本 (v3最新):通过浏览器本地离线运行,全新升级支持直接连接Supabase云端,实现测试数据上传、专属排行分享词、全网总榜与CSV批量上传,零部署成本! Python后端版本 (v3):通过Python后端发起请求,突破浏览器6并发限制,支持真正的高并发测试(50+并发),同步支持天梯榜。 本工具基于原作者纸鸢随风(B站)的作品进行魔改,增加了多项实用功能和优化,包括重试机制、图表自动显示、Prefill 和 Decode 速度分离、并发问题修复、请求超时控制、配置自动保存、并发吞吐修复、Token来源追踪等。 🔥 重要提示 ### 推荐使用 Python后端版本 🐍 Python后端版本功能更完善,维护更及时! #### ✅ Python后端版本优势(v3最新) - 🏆 独家天梯榜: 将测试结果上传至全网排行榜,获取专属分享链接 - 🚀 真实高并发:突破浏览器6并发限制,支持50+并发 - 📊 完整6图表体系:Prefill/Decode吞吐 + TTFT/ITL延迟 + 百分位统计 - 🔍 TTFT与ITL测量:首Token延迟 + Token间延迟(平均值+标准差/抖动) - 📈 全维度对比:历史记录对比支持6个图表,全方位性能分析 - 💾 增强导出:CSV包含全部数据,支持Excel无乱码,一键导出所有图表 - 🎯 更准确的吞吐计算:基于真实时间戳的并发总吞吐量 - ⚡ 积极维护:新功能优先在Python版本实现 #### ⚠️ 浏览器版本限制 - 并发限制:最多6并发(浏览器限制) - 测试准确性:最高并发下受浏览器网络调度限制较大 📖 快速开始: Python后端版本完整文档 💡 推荐流程: 1. 双击 python/启
| Stars | 97 |
| Forks | 9 |
| Language | HTML |
| Category | Agent Tool |
| Quality Score | 53.789323343007/100 |
| Open Issues | 1 |
| Last Updated | 2026-07-02 |
| Created | 2025-11-08 |
| Est. Tokens | ~19k |
Looking for a llm_speedtest alternative? If you're comparing llm_speedtest with other agent tool tools, these 4 projects are the closest alternatives on Agent Skills Hub — ranked by topic overlap, star count, and community traction.
Claude Code Skills for software engineering workflows - Git automation, testing, and code review
A Claude Code skill that turns PDFs, docs, and codebases into Obsidian study vaults
86 product management skills from Lenny's Podcast for Claude Code and AI agents. Hiring, user research, strate
Power rename/refactor tool (now with agent skill support!)
Explore other popular agent tool tools:
llm_speedtest is Local LLM Inference Speed Test Tool. It is categorized as a Agent Tool with 97 GitHub stars.
llm_speedtest is primarily written in HTML.
You can find installation instructions and usage details in the llm_speedtest GitHub repository at github.com/gengchaogit/llm_speedtest. The project has 97 stars and 9 forks, indicating an active community.
The top alternatives to llm_speedtest on Agent Skills Hub include claude-skills-marketplace, tutor-skills, lenny-skills. Each offers a different approach to the same problem space — compare them side-by-side by stars, quality score, and community activity.