reasoning benchmark
Terminal-Bench 2.1 Leaderboard
Terminal-Bench 2.1 is an updated benchmark that tests AI agents’ ability to operate a computer via the terminal.
Updated Sep 7, 2026
Models35
Model coverage35
MetricScore
EvidenceB
Terminal-Bench 2.1 Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelGO | Score89.40% | Percentile100.00% | Participants35 | EvidenceC | Evaluated |
| Rank02 | ModelOP | Score88.80% | Percentile97.06% | Participants35 | EvidenceC | Evaluated |
| Rank03 | ModelME | Score88.80% | Percentile94.12% | Participants35 | EvidenceC | Evaluated |
| Rank04 | ModelMA | Score88.30% | Percentile91.18% | Participants35 | EvidenceC | Evaluated |
| Rank05 | ModelZA | Score88.20% | Percentile88.24% | Participants35 | EvidenceC | Evaluated |
| Rank06 | ModelDE | Score87.90% | Percentile85.29% | Participants35 | EvidenceC | Evaluated |
| Rank07 | ModelOP | Score87.40% | Percentile82.35% | Participants35 | EvidenceC | Evaluated |
| Rank08 | ModelAC | Score86.60% | Percentile79.41% | Participants35 | EvidenceC | Evaluated |
| Rank09 | ModelGO | Score85.80% | Percentile76.47% | Participants35 | EvidenceC | Evaluated |
| Rank10 | ModelTE | Score85.40% | Percentile73.53% | Participants35 | EvidenceC | Evaluated |
| Rank11 | ModelOP | Score84.70% | Percentile70.59% | Participants35 | EvidenceC | Evaluated |
| Rank12 | ModelAN | Score84.30% | Percentile67.65% | Participants35 | EvidenceC | Evaluated |
| Rank13 | ModelZA | Score84.30% | Percentile64.71% | Participants35 | EvidenceC | Evaluated |
| Rank14 | ModelDE | Score83.90% | Percentile61.76% | Participants35 | EvidenceC | Evaluated |
| Rank15 | ModelXA | Score83.30% | Percentile58.82% | Participants35 | EvidenceC | Evaluated |
| Rank16 | ModelME | Score82.90% | Percentile55.88% | Participants35 | EvidenceC | Evaluated |
| Rank17 | ModelDE | Score82.70% | Percentile52.94% | Participants35 | EvidenceC | Evaluated |
| Rank18 | ModelZA | Score82.70% | Percentile50.00% | Participants35 | EvidenceC | Evaluated |
| Rank19 | ModelME | Score80.00% | Percentile47.06% | Participants35 | EvidenceC | Evaluated |
| Rank20 | ModelGO | Score78.00% | Percentile44.12% | Participants35 | EvidenceC | Evaluated |
| Rank21 | ModelAC | Score73.00% | Percentile41.18% | Participants35 | EvidenceC | Evaluated |
| Rank22 | ModelTE | Score71.70% | Percentile38.24% | Participants35 | EvidenceC | Evaluated |
| Rank23 | ModelBY | Score71.00% | Percentile35.29% | Participants35 | EvidenceC | Evaluated |
| Rank24 | ModelPO | Score70.20% | Percentile32.35% | Participants35 | EvidenceC | Evaluated |
| Rank25 | ModelBY | Score67.60% | Percentile29.41% | Participants35 | EvidenceC | Evaluated |
| Rank26 | ModelMI | Score66.00% | Percentile26.47% | Participants35 | EvidenceC | Evaluated |
| Rank27 | ModelTM | Score64.70% | Percentile23.53% | Participants35 | EvidenceC | Evaluated |
| Rank28 | ModelMI | Score62.90% | Percentile20.59% | Participants35 | EvidenceC | Evaluated |
| Rank29 | ModelUP | Score57.00% | Percentile17.65% | Participants35 | EvidenceC | Evaluated |
| Rank30 | ModelNV | Score56.40% | Percentile14.71% | Participants35 | EvidenceC | Evaluated |
Terminal-Bench 2.1 Highlights
The leading models and scores on this benchmark.
Terminal-Bench 2.1 Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for Terminal-Bench 2.1
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 2.1?
What Terminal-Bench 2.1 measures and how its scores work.
Terminal-Bench 2.1 is a benchmark for evaluating AI agents on real-world, end-to-end computer tasks performed through the terminal.
It measures how well models autonomously handle tasks including compiling code, training models, setting up servers, system administration, data science workflows, and security tasks, reported as a Score ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- Terminal-Bench 2.1
- 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 2.1.
Which model scores highest on Terminal-Bench 2.1?
Gemini 3.8 Flash is currently ranked first with 89.40%.
What are the top three models on Terminal-Bench 2.1?
The current leaders are Gemini 3.8 Flash (89.40%), GPT-5.6 Sol (88.80%), and Muse Spark 1.3 (88.80%).
Which Terminal-Bench 2.1 model has the lowest official input price?
GLM-5.3-Flash has the lowest matched official input price at $0.08 input / $0.25 output per 1M tokens.
Which models are fastest among Terminal-Bench 2.1 results?
The fastest matched records are GLM-5.3 (472.69 tok/s via FriendliAI), Gemini 3.7 Flash (117.86 tok/s via Google), 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.