reasoning benchmark
Terminal-Bench 4.0 Leaderboard
Terminal-Bench 4.0 evaluates AI agents on 66 real-world tasks in terminal and command-line containerized environments, including science-adjacent and frontier engineering problems.
Updated Sep 5, 2026
Models13
Model coverage13
MetricScore
EvidenceB
Terminal-Bench 4.0 Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelOP | Score57.70% | Percentile100.00% | Participants13 | EvidenceC | Evaluated |
| Rank02 | ModelAN | Score55.80% | Percentile91.67% | Participants13 | EvidenceC | Evaluated |
| Rank03 | ModelAN | Score51.80% | Percentile83.33% | Participants13 | EvidenceB | Evaluated |
| Rank04 | ModelAN | Score44.50% | Percentile75.00% | Participants13 | EvidenceB | Evaluated |
| Rank05 | ModelZA | Score41.80% | Percentile66.67% | Participants13 | EvidenceB | Evaluated |
| Rank06 | ModelOP | Score37.30% | Percentile58.33% | Participants13 | EvidenceB | Evaluated |
| Rank07 | ModelAN | Score23.60% | Percentile50.00% | Participants13 | EvidenceB | Evaluated |
| Rank08 | ModelOP | Score21.50% | Percentile41.67% | Participants13 | EvidenceB | Evaluated |
| Rank09 | ModelXA | Score20.30% | Percentile33.33% | Participants13 | EvidenceB | Evaluated |
| Rank10 | ModelGO | Score19.10% | Percentile25.00% | Participants13 | EvidenceC | Evaluated |
| Rank11 | ModelOP | Score17.30% | Percentile16.67% | Participants13 | EvidenceB | Evaluated |
| Rank12 | ModelAN | Score12.40% | Percentile8.33% | Participants13 | EvidenceB | Evaluated |
| Rank13 | ModelXA | Score12.40% | Percentile0.00% | Participants13 | EvidenceB | Evaluated |
Terminal-Bench 4.0 Highlights
The leading models and scores on this benchmark.
Terminal-Bench 4.0 Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for Terminal-Bench 4.0
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 4.0?
What Terminal-Bench 4.0 measures and how its scores work.
Terminal-Bench 4.0 is a reasoning benchmark for AI agents working in terminal and command-line containerized environments.
It measures performance across 66 tasks using the Score metric, reported as a ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- Terminal-Bench 4.0
- 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 4.0.
Which model scores highest on Terminal-Bench 4.0?
GPT-6 Astra is currently ranked first with 57.70%.
What are the top three models on Terminal-Bench 4.0?
The current leaders are GPT-6 Astra (57.70%), Claude Fable 5.1 (55.80%), and Claude Opus 5 (51.80%).
Which Terminal-Bench 4.0 model has the lowest official input price?
GPT-5.6 Luna has the lowest matched official input price at $0.20 input / $1.2 output per 1M tokens.
Which models are fastest among Terminal-Bench 4.0 results?
The fastest matched records are GLM-5.3 (472.69 tok/s via FriendliAI), Grok 4.6 (132.69 tok/s via xAI), and GPT-5.6 Terra (99.35 tok/s via OpenAI).
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.