Long-Horizon Terminal-Bench B
AI benchmark · Agentic
What is Long-Horizon Terminal-Bench?
Long-Horizon Terminal-Bench measures how well LLM agents sustain progress over hundreds of steps on 46 containerized terminal tasks across nine categories, graded with hidden verifiers that award partial credit.
“a terminal benchmark of 46 long-horizon tasks spanning nine categories, including experiment reproduction, software engineering, multimodal analysis, interactive games, and scientific computing.”
arxiv.org/abs/2607.08964 ↗“scientific computing, earth & energy systems, security & performance, research reproduction”
github.com/zli12321/LHTB ↗“Tasks span interactive games & puzzles, multimodal analysis, software / reverse engineering”
github.com/zli12321/LHTB ↗“scientific computing, earth & energy systems, security & performance, research reproduction”
github.com/zli12321/LHTB ↗“Tasks span interactive games & puzzles, multimodal analysis, software / reverse engineering”
github.com/zli12321/LHTB ↗
Who built Long-Horizon Terminal-Bench
Led by Tencent HY LLM Frontier.
Paper: arxiv.org/abs/2607.08964 ↗
Lineage
Benchmarks this one is built on and built from: versions, subsets, extensions and the indexes that combine them.
Built on
Terminal-Bench original
“Long-Horizon Terminal-Bench measures how well LLM agents sustain progress over hundreds of steps on 46 containerized terminal tasks across nine categories, graded with hidden verifiers that award partial credit.”
Tier B
Established: 15 or more models scored, or curated RL directories rate it.
How much stands behind a tier: usage, who backs it, a confirmed line from its own site or paper, and a clear read on X. High with three of the four, medium with two, low otherwise.
- 21 models scored
- in curated RL directories (1 pt)
On X · unclear, low buzz, from 6 posts about Long-Horizon Terminal-Bench in the last month
The only mentions are repeated posts from one account about Terminal-RSI gains on Terminal-Bench 2.1, with no direct discussion of Long-Horizon Terminal-Bench itself.
How we got to tier B
- Adoption: 21 models scored. Tier B.
- On X: unclear, low buzz. “The only mentions are repeated posts from one account about Terminal-RSI gains on Terminal-Bench 2.1, with no direct discussion of Long-Horizon Terminal-Bench itself.” No move.
- In curated RL directories: 1 point, a floor of none.
- Confirmed: a line from its own site or paper places it.
A clear view on X moves a tier one step up or down; most entries don't move. How tiers work.
Where Long-Horizon Terminal-Bench sits
Its domain on the map, live. Hover a logo for its name; click to open it. Open it on the full map
Next: browse all benchmarks. Entry from the RL Research daily scrape of public sources, 2026-11-05 snapshot.