Claude Code vs DeepSeek Harness
قارن النماذج
DeepSeek Harness is more worth it — 91% vs 60%
How we score worth-it? →Highest quality — but hits rate limits mid-session
- Because Claude gets the work done in one pass. It's like the other AIs are Internet Explorer and Claude is Chrome — once you try it, it's hard to forget.— v2ex
- Try Claude Code. You'll literally be able to automate 90% of the coding part of your job.— hn
- I'm ready to retire... Claude Code is giving me that same energy and drive. I'm chasing the midnight hour and not getting any sleep.— hn ↗
- I burned out a couple of times and assumed I will never finish so many sideprojects. Now I can just feed them into claude and guide it to completion. It feels great.— hn ↗
- I've never built anything outside of a python notebook before, but Claude Code felt like magic to me.— hn ↗
- Claude Code fabricates answers from search results when WebFetch fails, instead of reporting failure— github
- Claude Code really burns through quota — one task gets split into several requests.— v2ex
- Claude Code was the best harness from roughly release to January this year. Ever since then, it's become more and more bloated with no coherent plan or vision.— hn
- I've caught Claude making the gravest anti pattern mistakes using Elixir and trying to get it to correct them makes the whole thing worse.— hn ↗
- I really dislike Claude Code. The UX is terrible - it doesn't even have a per-turn diff feature. The model loves to act on its own and follows instructions very poorly. And token consumption is staggering.— v2ex ↗
“Best code I've ever had an AI write. It genuinely understands the codebase.”
“The pair programmer you actually want — it reads the whole repo before touching anything.”
“But the rate limits. Hit one mid-refactor and you're staring at a wall for an hour.”
Open-source agent framework — everything is a plugin
- I just installed DeepSeek Harness with the latest Bun version and am using it with a local 9B, speculative decoding Qwen 3.x variant, running in llama.cpp and it works GREAT for small python projects, so far. It was very easy to connect the harness to the local model and it seems to run quite fas...— hn
- the "everything is a plugin" claim is literal: model adapter, tool registry, session log and the agent loop itself are plugins. One consequence we liked: since plugins are just Cordis bundles, the same registry can be exposed to any MCP-speaking agent.— hn
- I really like the json schemas around the tool calls. Much stricter validation than in codex.— hn
- Deepseek Harness supposedly has nice plugin system which others do mostly lack.— hn
- Every Run is Traceable: everything the model sees is recorded in an append-only session log. Seems pretty helpful.— hn
- For all the high-powered theory it looked just like every other harness! I was expecting more.— hn
- «It uses an architecture where everything is a plugin» Ok, that's enough for me. I have developped over the year a plugin fatigue. Every product relying on "community plugins" for their features implies it works fine the 6 first months, then it's a nightmare of incompatible, deprecated plugins, w...— hn
- I wouldn't recommend beginners go messing around with DeepSeek Harness right now.— bilibili ↗
“Open-source agent framework where everything is a plugin.”
“Plugins, skills, MCP — you can wire up anything.”
“Great for self-hosted agent workflows and tinkerers.”
FAQ
Is Claude Code or DeepSeek Harness more worth it?
DeepSeek Harness is more worth it, with a 91% worth-it score vs 60% for Claude Code.
How much does Claude Code cost vs DeepSeek Harness?
Claude Code: Included in Pro/Max/Ultra · DeepSeek Harness: Free (open-source).
What do users say about Claude Code vs DeepSeek Harness?
Claude Code: Best code I've ever had an AI write. It genuinely understands the codebase.. DeepSeek Harness: Open-source agent framework where everything is a plugin..