Multi-step SWE
Multi-step SWE is 69 tasks · 574 dependent steps · 9 models. The strongest model solves 18.5% of tasks.. It is ranked #620 on the Dev Radar, in AI dev tools, first seen 4 h ago and shared in 1 post (421 views).
Multi-step SWE: Can Coding Agents Keep Building on Their Own Code Over Long Horizons? Skip to content Multi-step SWE A LONG-HORIZON CODING BENCHMARK Multi-step SWE Can Coding Agents Keep Building on Their Own Code Over Long Horizons? Paper ↗ GitHub ↗ 69 tasks 574 steps 9 languages 5–15 steps per task A step is a complete engineering request, not an agent turn or a tool call. 01 / LEADERBOARD Leaderboard Model + harness Tasks solved ↓ Steps passed Pass@4 Enable JavaScript to view the interactive results. 02 / A CLOSER LOOK Analysis Rollouts still on track Progress curve axis Task progress Step number Progress curve model 03 / EXPLORE THE…
What people said about Multi-step SWE on X
New multi-step harbor benchmark! 👏
— @harborframework, 4 h ago · 7 likes · see the post
Alternatives to Multi-step SWE
- flashprove — Further progress in provable inference will likely be an important part of improving trust in AI deployment. To that…
- Plugin Catalog — Open a pull request that adds your plugin to the Hermes plugin catalog
- niklaslenz.com — Play the world in your browser. Every number, the exact command and the raw output
- Agents in flight — Setup guide
- Introducing Unsloth Studio — Run and train AI models locally with Unsloth Studio.
- strata — Strata: how it works
Multi-step SWE in numbers
- Rank on the Dev Radar: #620 of 9317
- Shared in 1 post by 1 account: @harborframework
- 421 views on those posts
- First seen 4 h ago, last shared 4 h ago
- Pricing seen by AGK Intelligence: not stated
- Market: AI dev tools
FAQ
What is Multi-step SWE?
69 tasks · 574 dependent steps · 9 models. The strongest model solves 18.5% of tasks. It was first shared on X 4 h ago and is ranked #620 on the Dev Radar.
Is Multi-step SWE free?
Pricing is not stated on the page we read.
Who shared Multi-step SWE?
1 account on X, including @harborframework, in 1 post totalling 421 views.
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. 31.6k posts from 5k X accounts over the last 21 days, 9.3k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-05 00:17 UTC. Full method.