Benchmarks · coding
Terminal-Bench v4-AA
AA's 66-task Terminal-Bench v4.0, three repeats, average pass@1. Distinct from the upstream leaderboard and v2.1.
101 models · 185 results · data as of 2026-11-06 · artificialanalysis.ai ↗
Leaderboard
Kimi K3.1 Pro
Heavy line: best score at each price. Thin lines join one model's effort levels. Cost uses each model's median subscription plan, or API price with no plan data. 26 more models under All models in settings.
More from this benchmark
4 published measuresScore vs any published measure4 measures: sub-scores, splits, counts
76 more models under All models in settings.
Breakdown
darker is better within each column · click a column to sort| Model | ||||
|---|---|---|---|---|
| Kimi K3.1 max | 1,021,595,464 | 2,165,057 | 5,472,641 | 1,010,335,559 |
| Gemini 4 Carbon high | 948,085,121 | 4,311,276 | 6,294,617 | 928,860,538 |
| Sonnet 5.5 max | 4,108,901,435 | 6,131,049 | 19,109,282 | 4,076,088,441 |
| Opus 5.5 max | 1,958,776,924 | 4,386,279 | 12,308,317 | 1,936,161,085 |
| GPT-6 Astra xhigh | 167,415,558 | 1,138,247 | 1,717,925 | 161,855,952 |
| Gemini 4 Argon high | 1,002,215,176 | 4,109,435 | 6,802,366 | 982,942,792 |
| GPT-6.1 Sol max | 221,196,921 | 1,719,392 | 5,420,836 | 211,435,862 |
| GPT-6.1 Terra max | 861,905,441 | 2,350,716 | 4,343,651 | 849,634,684 |
| Fable 5.1 xhigh | 1,042,116,391 | 3,636,742 | 6,814,500 | 1,026,627,539 |
| Muse Wonder (Max) | 2,291,471,049 | 3,056,172 | 7,983,642 | 2,271,685,170 |
| Opus 5 max | 1,592,386,573 | 4,185,820 | 9,351,434 | 1,573,557,597 |
| GPT-6 Sol max | 808,817,407 | 2,149,309 | 4,277,242 | 797,482,117 |
| Fable 5 max, fallback | 1,298,495,457 | 4,004,881 | 10,299,536 | 1,280,283,215 |
| GLM 5.3 max | 1,717,065,779 | 3,081,614 | 9,759,733 | 1,699,352,466 |
| GPT-5.6 Sol max | 878,883,219 | 2,473,980 | 2,603,692 | 866,995,492 |
| Hy4 | 927,917,412 | 2,676,277 | 5,464,949 | 913,985,356 |
| GPT-5.6 Terra max | 1,293,055,632 | 3,198,765 | 5,499,792 | 1,276,021,076 |
| MiMo V2.6 Pro | 1,751,612,282 | 3,934,073 | 8,242,000 | 1,735,178,874 |
| DeepSeek V4.1 Flash 1017 max | 2,810,423,817 | 4,104,044 | 10,298,017 | 2,786,327,426 |
| Ling 3.1 Flash | 3,980,903,555 | 7,417,157 | 16,239,731 | 3,951,628,695 |
| Muse Spark 1.3 max | 1,843,746,742 | 4,188,655 | 9,587,438 | 1,823,226,206 |
| Haiku 5.5 max | 2,567,685,360 | 5,069,711 | 17,820,800 | 2,540,413,456 |
| GLM 5.3 Flash max | 1,660,201,056 | 2,920,298 | 9,029,262 | 1,643,733,656 |
| DeepSeek V4.1 Flash max | 3,805,240,530 | 4,828,883 | 13,997,756 | 3,777,962,649 |
| Grok 4.7 xhigh | 1,109,464,895 | 3,618,361 | 6,816,118 | 1,093,207,422 |
All results
185 rows| # | Model | Effort | Terminal-Bench v4-AA | Subscription $ / task | API $ / task | Harness |
|---|---|---|---|---|---|---|
| 1 | Kimi K3.1 | max | 67.2% | $5.27 | $5.27 | mini-SWE-agent |
| 2 | Gemini 4 Carbon | high | 65.2% | $3.58 | $10 | mini-SWE-agent |
| 3 | Sonnet 5.5 | max | 63.6% | $0.19 | $13 | mini-SWE-agent |
| 4 | Gemini 4 Carbon | low | 62.6% | $3.05 | $8.57 | mini-SWE-agent |
| 5 | Gemini 4 Carbon | medium | 60.6% | $3.14 | $8.82 | mini-SWE-agent |
| 6 | Opus 5.5 | max | 59.6% | $0.20 | $13 | mini-SWE-agent |
| 7 | Opus 5.5 | xhigh | 59.6% | $0.13 | $8.78 | mini-SWE-agent |
| 8 | GPT-6 Astra | xhigh | 59.6% | $0.25 | $5.86 | mini-SWE-agent |
| 9 | GPT-6 Astra | max | 59.1% | $0.36 | $8.50 | mini-SWE-agent |
| 10 | Sonnet 5.5 | xhigh | 57.1% | $0.092 | $6.08 | mini-SWE-agent |
| 11 | Gemini 4 Argon | high | 57.1% | $9.66 | $14 | mini-SWE-agent |
| 12 | Opus 5.5 | high | 56.6% | $0.077 | $5.12 | mini-SWE-agent |
| 13 | GPT-6.1 Sol | max | 56.1% | $0.11 | $1.82 | mini-SWE-agent |
| 14 | GPT-6.1 Terra | max | 55.6% | $0.075 | $1.49 | mini-SWE-agent |
| 15 | Fable 5.1 | xhigh | 55.1% | $2.42 | $16 | mini-SWE-agent |
| 16 | GPT-6 Astra | high | 54.0% | $0.17 | $4.05 | mini-SWE-agent |
| 17 | GPT-6.1 Sol | xhigh | 54.0% | $0.060 | $1.03 | mini-SWE-agent |
| 18 | Opus 5.5 | medium | 52.5% | $0.061 | $4.04 | mini-SWE-agent |
| 19 | Muse Wonder (Max) | max | 52.0% | $8.14 | $12 | mini-SWE-agent |
| 20 | Fable 5.1 | high | 52.0% | $1.79 | $11 | mini-SWE-agent |
| 21 | Fable 5.1 | max | 52.0% | $2.93 | $19 | mini-SWE-agent |
| 22 | GPT-6.1 Sol | high | 51.5% | $0.048 | $0.83 | mini-SWE-agent |
| 23 | GPT-6 Astra | medium | 49.5% | $0.19 | $4.43 | mini-SWE-agent |
| 24 | Opus 5 | max | 49.0% | $0.33 | $19 | mini-SWE-agent |
| 25 | GPT-6.1 Sol | medium | 48.0% | $0.036 | $0.61 | mini-SWE-agent |
| 26 | Opus 5 | xhigh | 46.5% | mini-SWE-agent | ||
| 27 | GPT-6.1 Terra | xhigh | 46.5% | $0.033 | $0.66 | mini-SWE-agent |
| 28 | Opus 5 | high | 46.0% | mini-SWE-agent | ||
| 29 | Fable 5.1 | medium | 44.9% | $1.40 | $8.98 | mini-SWE-agent |
| 30 | Sonnet 5.5 | high | 43.9% | $0.033 | $2.19 | mini-SWE-agent |