Benchmarks · agentic
τ³-Banking
τ³-Bench banking domain: multi-turn tool-using agent tasks with a simulated user; pass rate.
81 models · 135 results · data as of 2026-10-29 · artificialanalysis.ai ↗
Leaderboard
Grok 4.6 high
More from this benchmark
4 published measuresScore vs any published measure4 measures: sub-scores, splits, counts
Breakdown
darker is better within each column · click a column to sort| Model | ||||
|---|---|---|---|---|
| Grok 4.6 high | 223,199,893 | 271,025 | 725,954 | 211,127,195 |
| Muse Spark 1.3 max | 220,492,546 | 433,020 | 876,571 | 212,072,561 |
| GLM 5.3 max | 97,164,948 | 246,937 | 585,986 | 91,128,280 |
| Qwen 3.8 2.4T-A95B | 176,092,113 | 271,047 | 697,680 | 166,758,232 |
| Qwen 3.8 27B xhigh | 174,388,284 | 287,739 | 663,169 | 163,907,148 |
| Fable 5.1 max, fallback | 184,373,670 | 463,764 | 817,825 | 174,584,226 |
| GLM 5.3 Flash max | 109,485,999 | 251,107 | 876,436 | 102,690,607 |
| Kimi K3 max | 186,181,877 | 339,213 | 613,131 | 176,018,670 |
| Gemini 3.8 Flash medium | 193,079,993 | 264,978 | 1,293,412 | 184,926,397 |
| Opus 5 high | 140,871,245 | 428,528 | 440,299 | 131,931,139 |
| GPT-5.6 Sol max | 152,870,950 | 233,300 | 598,528 | 144,354,495 |
| GPT-6 Astra xhigh | 101,158,757 | 213,112 | 294,593 | 94,673,688 |
| Grok 4.5 high | 166,158,355 | 271,718 | 417,610 | 157,039,057 |
| GPT-5.6 Terra max | 141,126,363 | 212,197 | 762,471 | 133,028,016 |
| GPT-5.4 xhigh | 232,905,320 | 280,312 | 1,706,592 | 212,352,600 |
| GPT-5.5 xhigh | 117,652,838 | 206,711 | 643,558 | 106,325,646 |
| Ling 3.0 Flash Fin | 201,296,570 | 432,128 | 1,033,995 | 189,687,069 |
| Sonnet 5 max | 212,691,482 | 438,184 | 2,079,412 | 198,759,795 |
| Gemini 3.7 Flash medium | 121,016,382 | 226,667 | 631,755 | 115,002,101 |
| Muse Spark 1.2 xhigh | 112,669,880 | 387,308 | 476,936 | 107,497,897 |
| Opus 4.7 max | 163,541,617 | 437,858 | 1,109,879 | 153,225,753 |
| GLM 5.2 max | 97,751,708 | 245,028 | 1,842,488 | 90,587,768 |
| Opus 4.8 max | 136,754,829 | 469,832 | 2,711,406 | 125,581,055 |
| Gemini 3.5 Flash high | 235,769,397 | 288,264 | 1,529,476 | 226,189,946 |
| Muse Spark 1.1 xhigh | 171,369,967 | 469,231 | 577,154 | 162,406,683 |
All results
135 rows| # | Model | Effort | τ³-Banking |
|---|---|---|---|
| 1 | Grok 4.6 | high | 50.7% |
| 2 | Muse Spark 1.3 | max | 50.5% |
| 3 | GLM 5.3 | max | 50.3% |
| 4 | Qwen 3.8 2.4T-A95B | 49.1% | |
| 5 | Qwen 3.8 27B | xhigh | 48.0% |
| 6 | Qwen 3.8 27B | medium | 47.4% |
| 7 | Fable 5.1 | max | 47.2% |
| 8 | GLM 5.3 Flash | max | 47.2% |
| 9 | Muse Spark 1.3 | xhigh | 47.2% |
| 10 | Kimi K3 | max | 46.0% |
| 11 | Fable 5.1 | xhigh | 45.8% |
| 12 | Gemini 3.8 Flash | medium | 45.8% |
| 13 | Gemini 3.8 Flash | high | 44.9% |
| 14 | Opus 5 | high | 44.7% |
| 15 | GPT-5.6 Sol | max | 44.3% |
| 16 | Grok 4.6 | medium | 44.3% |
| 17 | Opus 5 | xhigh | 43.3% |
| 18 | Grok 4.6 | xhigh | 43.3% |
| 19 | Fable 5.1 | high | 43.1% |
| 20 | GPT-6 Astra | xhigh | 43.1% |
| 21 | Opus 5 | max | 42.1% |
| 22 | Grok 4.5 | high | 42.1% |
| 23 | Kimi K3 | low | 41.6% |
| 24 | GPT-6 Astra | max | 41.4% |
| 25 | Fable 5.1 | medium | 41.0% |
| 26 | GPT-5.6 Terra | max | 40.2% |
| 27 | GPT-6 Astra | high | 40.0% |
| 28 | GPT-5.4 | xhigh | 39.6% |
| 29 | Fable 5.1 | low | 39.0% |
| 30 | GPT-5.5 | xhigh | 39.0% |