本地模型替代 Claude — HN 讨论提炼
原文概要
HN 用户发帖问: 有没有人用本地模型完全替代了 Claude/GPT 做日常编码? 351 条评论中大量用户分享了实际配置. 主流方案: Qwen 3.6 系列 + Pi/opencode 工具链.
来源: HN 热门榜 (/best)
讨论焦点
1. 主流本地配置
“I replaced a $100/m subscription to claude in favor of running pi harness pointed at unsloth studio, using Qwen3.6-35B-A3B-MTP-GGUF and Gemma 4 31B.” — horsawlarway （用 pi + unsloth + Qwen3.6 系列替代了每月 $100 的 Claude.）
“Llama.cpp + Qwen3.6-35b (MTP) + OpenCode is quite capable and runs on a single RTX 3090 and is faster than most cloud models. Quality is like running edge models from 8-12 months ago.” — pierotofy （单张 RTX 3090 跑 Qwen3.6 系列 + OpenCode, 质量约等于 8-12 月前的边缘模型, 但速度快于云端.）
“I use pi with an RTX Pro 6000 Blackwell to run Gemma 4 31b for all my agentic coding.” — jodoherty
2. 硬件成本
“I’m using 4x RTX 5070’s and first-gen AMD threadripper to run Qwen3.6 27B (MTP) Q6_K with llama.cpp. Around 50-60 toks/sec.” — jborak
“Qwen3.6-35B-A3B on a Strix Halo 128GB. I have way too much VRAM for such a model but Qwen never released the 122B version of Qwen3.6.” — stymaar
3. 对话质量 vs 速度
“It is NOT as smart as CC or Codex but its enough to get most of my work done. I didn’t set out to replace Claude, I set out to save money.” — bluejay2387 （不如 Codex/Claude 聪明, 但足够完成大部分工作. 我不是想替代, 只是想省钱.）
“The problem with this question is that it encompasses a huge spectrum of capabilities and expectations. If you can only run an 8B model and expect it to be good at vibe coding, you’ll be disappointed.” — sosodev （这问题太宽泛. 只能跑 8B 模型还指望 vibe coding 的好效果会失望的.）
4. 隐私考量
“I care about data privacy and LLMs being free. I’m using the Pi coding harness but containerized and sandboxed, to make sure it’s running completely offline.” — Greenpants
“For client projects where privacy and security is important, but no enterprise contract: Open code against Infomaniak hosted OSS models.” — ozten
典型观点
| 立场 | 用户 | 一句话 |
|---|---|---|
| 🔵 可替代 | horsawlarway | Qwen3.6 + pi 已完全替代 $100/m 的 Claude |
| 🔵 省钱 | bluejay2387 | 不如云模型聪明, 但省了每月一百刀 |
| 🟡 硬件门槛 | sosodev | 跑 8B 模型期望 vibe coding 基本没戏 |
| 🟡 质量差距 | pierotofy | 本地模型质量约等于 8-12 月前的云模型 |
| ⚪ 隐私优先 | Greenpants | 容器化 + 沙箱, 完全离线运行 |
| ⚪ 主流组合 | 多数 | Qwen3.6 + Pi/OpenCode + llama.cpp 是默认方案 |