reasoning benchmark
Terminal-Bench Hard Leaderboard
Terminal-Bench Hard is a harder terminal-agent benchmark variant evaluated with the Terminus-2 harness in Cohere's Command A+ and North Mini Code releases.
Updated Sep 5, 2026
Models2
Model coverage2
MetricScore
EvidenceB
Terminal-Bench Hard Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelCO | Score31.10% | Percentile100.00% | Participants2 | EvidenceC | Evaluated |
| Rank02 | ModelCO | Score25.00% | Percentile0.00% | Participants2 | EvidenceC | Evaluated |
Terminal-Bench Hard Highlights
The leading models and scores on this benchmark.
Terminal-Bench Hard Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for Terminal-Bench Hard
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is Terminal-Bench Hard?
What Terminal-Bench Hard measures and how its scores work.
Terminal-Bench Hard is a harder terminal-agent benchmark variant.
It reports a Score measured as a ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- Terminal-Bench Hard
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about Terminal-Bench Hard.
Which model scores highest on Terminal-Bench Hard?
North Mini Code 1.0 is currently ranked first with 31.10%.
What are the top three models on Terminal-Bench Hard?
The current leaders are North Mini Code 1.0 (31.10%) and Command A+ (25.00%).
Which Terminal-Bench Hard model has the lowest official input price?
Command A+ has the lowest matched official input price at $2.5 input / $10 output per 1M tokens.
Which models are fastest among Terminal-Bench Hard results?
No matched runtime record is currently available.
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.