BREAKING: same open model, same provider, same reasoning budget. Swap only the coding…
This is a AI post classified by AGK Intelligence as AI coding agents (a model release), kept by the AI Radar because it carries real work, not commentary.
BREAKING: same open model, same provider, same reasoning budget. Swap only the coding harness and scores swing by ~24 points. Joël Niklaus just published the Harness Leaderboard: 28,849 trials, $4,746 of inference, nine harnesses on GLM-5.3-Flash, DeepSeek-V4.1-Flash, and Flint across Terminal-Bench 2.1 and DeepSWE. Averaged across those models, mini-swe-agent clears 51.4% of tasks. OpenClaw lands at 27.3%. Four other harnesses sit statistically tied with the leader. Pi hits 49.4% at $0.29 per solved task, about 45% cheaper than mini-swe-agent's $0.53. OpenCode runs $3.35 per solve, ~11.7x P
Posted by Guillermo Flor (14k followers) 1 h ago · 4 likes · 300 views · view the original post on X. Kept by the AI Radar as AI coding agents.
More AI work like this
- Qwen 4 just wrote a real time fluid simulator in one HTML file. First try 🔥 — @aijoey
- AI Daily Radar, Friday, Nov 6 (by SMF Works) — @MichaelGannotti
- Building the Dropbox experience meant changing how we build software. See how our… — @Dropbox
- Cursor was about to raise a few billion dollars. Then it sold to SpaceX for $60B. — @guilleflorvs
- The tricky part of sandboxing is often what happens at the boundary with the host. — @AgenticAIFdn
- does paying more for AI get you better 3D? — @DataChaz
- The Daytona Playbook: From $1M to $100M in a Year — @guilleflorvs
- Claude Code 2.1.318 对用 claude -p 挂定时任务的人很实用： — @wei_wang
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. 65.3k posts from 5k X accounts over the last 14 days, 10.4k tools, 19 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-06 17:49 UTC. Full method.