Terminal-Bench 4.0 (AA) leaderboard
As of 9 Nov 2026, the top score on Terminal-Bench 4.0 (AA) is 67.2% by Kimi K3.1 Pro (Moonshot AI), ahead of Gemini 4 Carbon at 65.2%, across 250 model configurations with a published result (Artificial Analysis, result dated 23 Oct 2026). Real command-line tasks in a sandboxed terminal, version 4.0. Run by Artificial Analysis; the Terminal-Bench team's own run is the one in the index. Terminal-Bench 4.0 (AA) is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 9 Nov 2026)
- Kimi K3.1 Pro holds the top Terminal-Bench 4.0 (AA) score at 67.2% (Artificial Analysis, 23 Oct 2026).
- Gemini 4 Carbon is second on Terminal-Bench 4.0 (AA) at 65.2% (Artificial Analysis, 28 Oct 2026).
- Terminal-Bench 4.0 (AA): Real command-line tasks in a sandboxed terminal, version 4.0. Run by Artificial Analysis; the Terminal-Bench team's own run is the one in the index (Artificial Analysis).
- 250 model configurations have a published Terminal-Bench 4.0 (AA) result on BenchLeader (BenchLeader, 9 Nov 2026).
- Terminal-Bench 4.0 (AA) does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
Real command-line tasks in a sandboxed terminal, version 4.0. Run by Artificial Analysis; the Terminal-Bench team's own run is the one in the index.
- Published by
- Artificial Analysis
- Category
- Coding
- Index weight
- Reference only
- Models
- 250
- Data as of
- 9 Nov 2026
Source: Artificial Analysis (artificialanalysis.ai). Data taken from the public leaderboard.
- 1Kimi K3.1 Pro67.2%
- 2Gemini 4 Carbon (high)65.2%
- 3Claude Sonnet 5.5 (max)63.6%
- 4Gemini 4 Carbon (low)62.6%
- 5Claude Fable 5.2 (xhigh)61.6%
- 6Gemini 4 Carbon (medium)60.6%
- 7Claude Fable 5.2 (max)59.6%
- 8Claude Opus 5.5 (max)59.6%
- 9Claude Fable 5.2 (high)59.6%
- 10GPT-6 Astra (xhigh)59.6%
- 11Claude Opus 5.5 (xhigh)59.6%
- 12GPT-6 Astra (max)59.1%
- 13Gemini 4 Argon (high)57.1%
- 14Claude Sonnet 5.5 (xhigh)57.1%
- 15Muse Wonder Full (max)57.1%
250 of 250
| # | ||||||
|---|---|---|---|---|---|---|
| 1 |  Kimi K3.1 Proopen ↗ | 67.2% | 61.2 | $5.27 | 116k | 2026-10-23 |
| 2 |  Gemini 4 Carbonhigh | 65.2% | 68.5 | $10.05 | 161k | 2026-10-28 |
| 3 |  Claude Sonnet 5.5max | 63.6% | 65.0 | $12.65 | 382k | 2026-09-28 |
| 4 | Gemini 4 Carbonlow | 62.6% | – | $8.57 | 139k | 2026-10-28 |
| 5 | NewClaude Fable 5.2xhigh | 61.6% | 68.7 | $20.81 | 260k | 2026-11-09 |
| 6 | Gemini 4 Carbonmedium | 60.6% | 64.1 | $8.82 | 143k | 2026-10-28 |
| 7 | NewClaude Fable 5.2max | 59.6% | 69.4 | $31.59 | 377k | 2026-11-09 |
| 8 | Claude Opus 5.5max | 59.6% | 68.9 | $13.11 | 253k | 2026-09-22 |
| 9 | NewClaude Fable 5.2high | 59.6% | 68.9 | $13.47 | 158k | 2026-11-09 |
| 10 |  GPT-6 Astraxhigh | 59.6% | 66.2 | $5.86 | 43k | 2026-09-03 |
| 11 | Claude Opus 5.5xhigh | 59.6% | 65.5 | $8.78 | 169k | 2026-09-22 |
| 12 | GPT-6 Astramax | 59.1% | 67.4 | $8.50 | 76k | 2026-09-03 |
| 13 | Gemini 4 Argonhigh | 57.1% | 65.7 | $13.56 | 165k | 2026-09-30 |
| 14 | Claude Sonnet 5.5xhigh | 57.1% | 62.7 | $6.08 | 192k | 2026-09-28 |
| 15 |  Muse Wonder Fullmax | 57.1% | 62.2 | – | 153k | 2026-10-12 |
| 16 | Claude Opus 5.5high | 56.6% | 65.4 | $5.12 | 100k | 2026-09-22 |
| 17 | GPT-6.1 Solmax | 56.1% | 67.0 | $1.82 | 108k | 2026-09-29 |
| 18 | GPT-6.1 Terramax | 55.6% | 65.0 | $1.49 | 101k | 2026-10-26 |
| 19 | NewClaude Fable 5.2medium | 55.0% | 66.3 | $5.25 | 63k | 2026-11-09 |
| 20 | Claude Fable 5.1xhigh | 55.0% | 64.6 | $15.53 | 158k | 2026-09-01 |
| 21 | NewClaude Fable 5.2low | 54.5% | 65.3 | $4.34 | 53k | 2026-11-09 |
| 22 | GPT-6 Astrahigh | 54.0% | 66.5 | $4.05 | 30k | 2026-09-03 |
| 23 | GPT-6.1 Solxhigh | 54.0% | 63.8 | $1.03 | 57k | 2026-09-29 |
| 24 | Claude Opus 5.5medium | 52.5% | 61.5 | $4.04 | 76k | 2026-09-22 |
| 25 | Claude Fable 5.1max | 52.0% | 66.9 | $18.77 | 192k | 2026-09-01 |
| 26 | Claude Fable 5.1high | 52.0% | 63.3 | $11.49 | 113k | 2026-09-01 |
| 27 | Muse Wondermax | 52.0% | 61.9 | $12.21 | 167k | 2026-10-12 |
| 28 | GPT-6.1 Solhigh | 51.5% | 63.5 | $0.828 | 44k | 2026-09-29 |
| 29 | GPT-6 Astramedium | 49.5% | 64.5 | $4.43 | 30k | 2026-09-03 |
| 30 | Claude Opus 5max | 49.0% | 65.1 | $19.29 | 205k | 2026-07-24 |
| 31 | GPT-6.1 Solmedium | 48.0% | 62.9 | $0.614 | 32k | 2026-09-29 |
| 32 | Claude Opus 5xhigh | 46.5% | 60.1 | – | – | 2026-07-24 |
| 33 | GPT-6.1 Terraxhigh | 46.5% | 60.1 | $0.657 | 47k | 2026-10-26 |
| 34 | Claude Opus 5high | 46.0% | 64.2 | – | – | 2026-07-24 |
| 35 | Claude Fable 5.1medium | 45.0% | 61.7 | $8.98 | 84k | 2026-09-01 |
| 36 | GPT-6 Solmax | 43.9% | 62.9 | $4.00 | 97k | 2026-09-22 |
| 37 | Claude Sonnet 5.5high | 43.9% | 58.0 | $2.19 | 77k | 2026-09-28 |
| 38 | Claude Fable 5max | 42.4% | 65.3 | – | – | 2026-06-09 |
| 39 | GPT-6 Astralow | 41.9% | 62.2 | $2.25 | 15k | 2026-09-03 |
| 40 | GPT-6.1 Terrahigh | 41.9% | 59.7 | $0.438 | 30k | 2026-10-26 |
| 41 |  GLM 5.3maxopen ↗ | 41.9% | 58.9 | $8.06 | 195k | 2026-08-18 |
| 42 | NewMuse Embermax | 40.9% | – | $0.266 | 148k | 2026-11-09 |
| 43 | Claude Fable 5.1low | 40.4% | 58.7 | $7.20 | 66k | 2026-09-01 |
| 44 | GPT-5.6 Solmax | 39.9% | 65.4 | – | – | 2026-07-09 |
| 45 |  Qwen3.8 Maxmax | 38.9% | 58.9 | $18.68 | 293k | 2026-09-02 |
| 46 | Muse Wonderhigh | 38.4% | 56.3 | $8.10 | 116k | 2026-10-12 |
| 47 |  Hy4open ↗ | 37.4% | 55.3 | $1.64 | 123k | 2026-11-03 |
| 48 | GPT-5.6 Terramax | 35.4% | 60.5 | $6.39 | 132k | 2026-07-09 |
| 49 | NewMuse Emberxhigh | 35.4% | – | $0.222 | 132k | 2026-11-09 |
| 50 |  MiMo-V2.6-Proopen ↗ | 34.9% | 57.1 | $0.438 | 184k | 2026-09-21 |
| 51 | GLM 5.3lowopen ↗ | 34.9% | 47.4 | $6.40 | 128k | 2026-08-18 |
| 52 | Claude Opus 5medium | 34.3% | 57.4 | – | – | 2026-07-24 |
| 53 |  DeepSeek V4.1 Flash 1017maxopen ↗ | 33.8% | 58.9 | $0.880 | 218k | 2026-10-17 |
| 54 | Muse Spark 1.3max | 33.3% | 61.7 | $5.58 | 209k | 2026-09-02 |
| 55 |  Step 5 Previewopen | 33.3% | 54.6 | $5.18 | 260k | 2026-09-18 |
| 56 |  Ling 3.1 Flashopen | 33.3% | 50.8 | – | – | 2026-10-01 |
| 57 | GLM 5.3 Flashopen ↗ | 32.8% | 56.8 | $0.889 | 181k | 2026-08-26 |
| 58 | Claude Haiku 5.5max | 32.8% | 56.8 | $0.652 | 347k | 2026-10-07 |
| 59 | GPT-6.1 Terramedium | 32.3% | 58.6 | $0.276 | 19k | 2026-10-26 |
| 60 | Claude Opus 5.5low | 31.3% | 59.0 | $2.08 | 38k | 2026-09-22 |
More coding benchmarks · All LLM benchmarks
Related benchmarks
Top head-to-heads on Terminal-Bench 4.0 (AA)
Cite as: BenchLeader, “Terminal-Bench 4.0 (AA) leaderboard”, data as of 9 Nov 2026, https://www.benchleader.com/benchmarks/aa_terminalbench_40.
Terminal-Bench 4.0 (AA): questions
- What does Terminal-Bench 4.0 (AA) measure?
- Real command-line tasks in a sandboxed terminal, version 4.0. Run by Artificial Analysis; the Terminal-Bench team's own run is the one in the index. Scores are reported in percent of tasks solved; higher is better.
- What is the top score on Terminal-Bench 4.0 (AA)?
- As of 9 Nov 2026, the highest score on Terminal-Bench 4.0 (AA) is 67.2% by Kimi K3.1 Pro (result dated 23 Oct 2026), ahead of Gemini 4 Carbon at 65.2%. Scores come from Artificial Analysis and are re-read daily.
- Which AI model leads Terminal-Bench 4.0 (AA)?
- Kimi K3.1 Pro leads Terminal-Bench 4.0 (AA) with 67.2% as of 9 Nov 2026 (result dated 23 Oct 2026), ahead of Gemini 4 Carbon at 65.2%.
- How many models have Terminal-Bench 4.0 (AA) results?
- 250 model configurations have a Terminal-Bench 4.0 (AA) result on BenchLeader, all taken from Artificial Analysis.
- Who runs Terminal-Bench 4.0 (AA) and how often is it updated?
- Terminal-Bench 4.0 (AA) is published by Artificial Analysis. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does Terminal-Bench 4.0 (AA) count toward the BenchLeader Index?
- No. Terminal-Bench 4.0 (AA) is shown for reference but left out of the composite index.