Score breakdown
Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.
Why it matters
Useful for eval teams measuring long-horizon agent endurance (hundreds of steps, stateful terminal environments) rather than single-artifact coding, for teams benchmarking agents against hidden verifiers that resist self-reported progress, and for anyone building on Terminal-Bench / Harbor who needs the long-horizon companion benchmark.
Where this stands now
zli12321/LHTB ranks #1065 of 3275 tracked Radar items by composite score (7.0 against a section median of 4.9). The section currently carries 2159 Bronze, 645 Gold, 471 Silver. RepoRadar has retained observations for this record since 2026-08-21 (52 days in the current window). Signal extremes versus the section: momentum at the 76th percentile; novelty at the 72th percentile.
Who should use it
Who should skip it
Skip zli12321/LHTB if the source repository or demo is inactive, unmaintained, or no longer matches the description shown here.
About this signal
zli12321/LHTB is tracked by RepoRadar as an AI project in the Radar section. First seen 2026-08-21; the source record was last checked on 2026-08-21. The current verdict is 'try now' with a Silver tier and moderate setup difficulty. Across RepoRadar's eight signals, zli12321/LHTB is strongest on open-source/build quality (8.4) and workflow potential (8.1) and weakest on setup ease (6.0) — a profile worth weighing against your own priorities. This page summarizes the evidence RepoRadar captured from https://github.com/zli12321/LHTB.
How this item is evaluated
The zli12321/LHTB record combines a 7.0/10 composite score with separate popularity (100.0), risk (low), and setup (moderate) signals. See the scoring methodology for the current weights and evidence definitions.
Questions worth asking before you adopt this
Putting this into practice? Read How to read AI benchmarks without getting fooled for the checklist behind this score.
Risk explanation
Benchmark runs require the modified Harbor bundled in the repo: stock upstream Harbor ignores the continue-until-timeout flag, so those tasks run single-shot there and score lower -- numbers are NOT comparable across harnesses, and any reported LHTB score must state the harness; Diagnostic-feedback runs (HB_VERIFIER_FEEDBACK_MODE=diagnostic) are explicitly not benchmark-comparable; the default binary mode is the only comparable configuration; The 46-task suite is small relative to large-scale benchmarks, so per-task variance is high and cross-model comparisons need multiple seeds/runs; The benchmark measures endurance in containerized terminal environments specifically; it does not measure breadth across non-terminal agent surfaces.