joelniklaus/harness-leaderboard
docker · ♥ 161
- Likes29
- Licenseapache-2.0
docker · ♥ 161
Which coding-agent harness should you use for your model?
Which coding-agent harness should you run your open model in? Same model, same provider, same reasoning effort, three attempts with confidence intervals, every trajectory public.
This is Joel Niklaus's leaderboard. It measures which coding-agent harness to run a given model in: Claude Code, Codex, Pi, OpenCode, Hermes Agent, mini-swe-agent, Kimi Code, Qwen Code and OpenClaw. The models so far are GLM-5.3-Flash and DeepSeek-V4.1-Flash (low and high reasoning effort) and Flint (default effort). The benchmarks are Terminal-Bench 2.1 and DeepSWE v1.1, run through Harbor on Hugging Face…
Excerpt from the Space card on Hugging Face ↗.
joelniklaus/harness-leaderboard was first seen on Hugging Face · Trending spaces on Nov 6, 2026. Since then it has appeared in 53 checks on 3 different days. Its best position was #2 of 30, reached on Nov 8, 14:42 UTC. It is #2 right now, climbing from #4 to #2 over the last 6 checks.
Position on Hugging Face · Trending spaces at every check (lower is better).
Best rank per day (UTC).
🔥 as reported by Hugging Face.
Archived daily, weekly and monthly lists that include this entry.
Yes. As of Nov 8, 14:42 UTC it is #2, climbing from #4 to #2 over the last 6 checks.
#2 on Hugging Face · Trending spaces, on Nov 8, 14:42 UTC. Trendoday checks this list every 60 minutes.
It was first seen on Nov 6, 2026 and has been on the list on 3 days (53 checks) since.