New Tsinghua and Microsoft paper finds that DeepSeek Harness used 87.93% fewer tokens…
This is a vibe coding post classified by AGK Intelligence as Coding agents (a workflow), kept by the Vibe Coding Radar because it carries real work, not commentary.
New Tsinghua and Microsoft paper finds that DeepSeek Harness used 87.93% fewer tokens than its original version on 98 HumanEval tasks while success rose from 71.43% to 85.71%. Agent tools like Codex and DeepSeek Harness call the LLM after each tool result to choose the next step. Often that step was already decided, like running tests once the code compiles. HarnessMM lets the model write future commands with exact arguments, plus a check for each, such as a successful build. When a tool result passes the check, the agent runs the command and skips the LLM call. On 98 HumanEval tasks, Codex
Posted by Rohan Paul (160.7k followers) 4 days ago · 267 likes · 13.5k views · view the original post on X. Kept by the Vibe Coding Radar as Coding agents.
More vibe coding work like this
- This week at Entelligence, we launched Entelligence Memory: one shared memory for every… — @EntelligenceAI
- ⬛ Something is opening. — @Thirdfy_
- Changing coding agents? Teams Plan in @DevinAI is 50% off for three months! — @Chris_Wozniczek
- Kimi K3.1 Pro is now live on @akashnetAI. — @akashnet
- Here's 4 steps to run an uncensored AI on your own GPU for $0: — @DeRonin_
- just farm, built in 5 weeks with opus 5.5 and unity 6 per the post. the dev's rule is… — @0xSpikez
- Can coding agents replace traditional 3D vision pipelines and build a entire interactive… — @EinsiaAI
- 🚀 Work across multiple agents, keep your favorite canvases close, and customize how you… — @code
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 7.9k posts from 5k X accounts over the last 14 days, 2.6k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-06 17:48 UTC. Full method.