TauBench V3 Banking

TauBench V3 - Banking domain. Agentic tool use benchmark for banking customer service.

Category

General Agents

Max Score

100.0

Score Type

percent

Active

Yes

Model Scores

Scores for TauBench V3 Banking

Model Score Date Verified
NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16
47.57%
04.06.2026 Verified
MiniMax-M2.7
8.74%
04.06.2026 Unverified
Kimi K2.6
50.00%
04.06.2026 Unverified
Qwen3.5-397B-A17B
39.32%
04.06.2026 Unverified
DeepSeek-V4-Pro
63.59%
04.06.2026 Unverified
DeepSeek-V4-Flash
67.48%
04.06.2026 Unverified
Muse-Glimmer-30B
51.94%
01.08.2026 Verified
Gemma4-31B-it
11.16%
01.08.2026 Unverified
Qwen3.6-27B
18.93%
01.08.2026 Unverified
Kimi K3
100.00%
18.08.2026 Verified
GLM-5.1
12.80
04.06.2026 Unverified