Agentic & Tool Use benchmark

# τ²-bench Banking leaderboard

> τ²-bench Banking results for 26 AI models, led by Qwen3.8 Max at 55.1%. What the benchmark measures, who runs it, and a source for every score.
- Canonical page: https://noometry.com/benchmarks/tau2-banking
- Last updated: 2026-10-11
- Title: τ²-bench Banking Leaderboard (October 2026): Scores by Model

As of October 2026, Qwen3.8 Max has the highest published τ²-bench Banking score on Noometry at 55.1%, out of 26 models with results.

Last verified October 11, 2026

## About τ²-bench Banking

Banking customer-service tasks that need the agent to search a knowledge base as well as use account tools under policy.

- **Category:** [Agentic & Tool Use](https://noometry.com/best/agentic)
- **Format:** Tool use with retrieval
- **Unit:** Percent (random guessing ≈ 0%)
- **Official site:** [taubench.com](https://taubench.com/)

## Top 15 models

Top models on τ²-bench Banking

1.  Qwen3.8 Max 55.1%
2.  Claude Opus 5 48.7%
3.  Grok 4.5 47.9%
4.  GPT-5.6 Sol 46.9%
5.  GPT-5.5 44.6%
6.  Muse Spark 1.1 40.5%
7.  Claude Opus 4.7 40.2%
8.  Claude Fable 5 39.7%
9.  Claude Opus 4.8 39.7%
10.  GPT-5.4 39.4%
11.  GLM-5.2 37.1%
12.  Kimi K3 37.1%
13.  GPT-5.2 32.2%
14.  Claude Opus 4.6 27.3%
15.  Gemini 3 Flash Preview 27.3%
16.  0204060

Sponsored placements are available on pages like this one. [Advertise on Noometry](https://noometry.com/advertise)

## All results

τ²-bench Banking results by model
| # | Model | Provider | Score | Setting | Source | Date |
| --- | --- | --- | --- | --- | --- | --- |
| 1 | [Qwen3.8 Max](https://noometry.com/models/qwen3-8-max) |  [![](/logos/alibaba.svg) Alibaba (Qwen)](https://noometry.com/providers/alibaba) | 55.1% | xhigh | [τ²-bench](https://taubench.com/) | 2026-08-04 |
| 2 | [Claude Opus 5](https://noometry.com/models/claude-opus-5) | [Anthropic](https://noometry.com/providers/anthropic) | 48.7% | max | [τ²-bench](https://taubench.com/) | 2026-08-04 |
| 3 | [Grok 4.5](https://noometry.com/models/grok-4-5) | [xAI](https://noometry.com/providers/xai) | 47.9% | high | [τ²-bench](https://taubench.com/) | 2026-08-04 |
| 4 | [GPT-5.6 Sol](https://noometry.com/models/gpt-5-6-sol) | [OpenAI](https://noometry.com/providers/openai) | 46.9% | xhigh | [τ²-bench](https://taubench.com/) | 2026-08-04 |
| 5 | [GPT-5.5](https://noometry.com/models/gpt-5-5) | [OpenAI](https://noometry.com/providers/openai) | 44.6% | xhigh | [τ²-bench](https://taubench.com/) | 2026-05-05 |
| 6 | [Muse Spark 1.1](https://noometry.com/models/muse-spark-1-1) |  [![](/logos/meta.svg) Meta](https://noometry.com/providers/meta) | 40.5% | xhigh | [τ²-bench](https://taubench.com/) | 2026-08-04 |
| 7 | [Claude Opus 4.7](https://noometry.com/models/claude-opus-4-7) | [Anthropic](https://noometry.com/providers/anthropic) | 40.2% | max | [τ²-bench](https://taubench.com/) | 2026-05-05 |
| 8 | [Claude Fable 5](https://noometry.com/models/claude-fable-5) | [Anthropic](https://noometry.com/providers/anthropic) | 39.7% | max | [τ²-bench](https://taubench.com/) | 2026-08-04 |
| 9 | [Claude Opus 4.8](https://noometry.com/models/claude-opus-4-8) | [Anthropic](https://noometry.com/providers/anthropic) | 39.7% | max | [τ²-bench](https://taubench.com/) | 2026-08-04 |
| 10 | [GPT-5.4](https://noometry.com/models/gpt-5-4) | [OpenAI](https://noometry.com/providers/openai) | 39.4% | xhigh | [τ²-bench](https://taubench.com/) | 2026-03-25 |
| 11 | [GLM-5.2](https://noometry.com/models/glm-5-2) | [Z.ai (Zhipu)](https://noometry.com/providers/zai) | 37.1% | xhigh | [τ²-bench](https://taubench.com/) | 2026-08-04 |
| 12 | [Kimi K3](https://noometry.com/models/kimi-k3) | [Moonshot AI](https://noometry.com/providers/moonshot) | 37.1% | max | [τ²-bench](https://taubench.com/) | 2026-08-04 |
| 13 | [GPT-5.2](https://noometry.com/models/gpt-5-2) | [OpenAI](https://noometry.com/providers/openai) | 32.2% | high | [τ²-bench](https://taubench.com/) | 2026-02-26 |
| 14 | [Claude Opus 4.6](https://noometry.com/models/claude-opus-4-6) | [Anthropic](https://noometry.com/providers/anthropic) | 27.3% | max | [τ²-bench](https://taubench.com/) | 2026-05-05 |
| 15 | [Gemini 3 Flash Preview](https://noometry.com/models/gemini-3-flash-preview) |  [![](/logos/google.svg) Google](https://noometry.com/providers/google) | 27.3% | high | [τ²-bench](https://taubench.com/) | 2026-03-02 |
| 16 | [Gemini 3.1 Pro Preview](https://noometry.com/models/gemini-3-1-pro-preview) |  [![](/logos/google.svg) Google](https://noometry.com/providers/google) | 26% | high | [τ²-bench](https://taubench.com/) | 2026-05-05 |
| 17 | [Claude Sonnet 4.5](https://noometry.com/models/claude-sonnet-4-5) | [Anthropic](https://noometry.com/providers/anthropic) | 25.3% | enabled | [τ²-bench](https://taubench.com/) | 2026-02-26 |
| 18 | [Inkling](https://noometry.com/models/inkling) | [T Thinking Machines Lab](https://noometry.com/providers/thinking-machines) | 25% | max | [τ²-bench](https://taubench.com/) | 2026-08-04 |
| 19 | [Claude Opus 4.5](https://noometry.com/models/claude-opus-4-5) | [Anthropic](https://noometry.com/providers/anthropic) | 24.7% | high | [τ²-bench](https://taubench.com/) | 2026-02-26 |
| 20 | [Gemini 3 Pro](https://noometry.com/models/gemini-3-pro) |  [![](/logos/google.svg) Google](https://noometry.com/providers/google) | 18% | high | [τ²-bench](https://taubench.com/) | 2026-03-02 |
| 21 | [Grok 4.20 (Non-Reasoning)](https://noometry.com/models/grok-4-20) | [xAI](https://noometry.com/providers/xai) | 18% | high | [τ²-bench](https://taubench.com/) | 2026-05-05 |
| 22 | [Grok 4 Fast](https://noometry.com/models/grok-4-fast) | [xAI](https://noometry.com/providers/xai) | 15.7% | high | [τ²-bench](https://taubench.com/) | 2026-05-05 |
| 23 | [Gemini 2.5 Pro](https://noometry.com/models/gemini-2-5-pro) |  [![](/logos/google.svg) Google](https://noometry.com/providers/google) | 13.7% | high | [τ²-bench](https://taubench.com/) | 2026-05-05 |
| 24 | [Grok 4.1 Fast](https://noometry.com/models/grok-4-1-fast) | [xAI](https://noometry.com/providers/xai) | 13.1% | high | [τ²-bench](https://taubench.com/) | 2026-05-05 |
| 25 | [GLM-5](https://noometry.com/models/glm-5) | [Z.ai (Zhipu)](https://noometry.com/providers/zai) | 9.8% | enabled | [τ²-bench](https://taubench.com/) | 2026-03-02 |
| 26 | [Qwen3.5 397B-A17B](https://noometry.com/models/qwen3-5-397b-a17b) |  [![](/logos/alibaba.svg) Alibaba (Qwen)](https://noometry.com/providers/alibaba) | 9.8% | enabled | [τ²-bench](https://taubench.com/) | 2026-03-02 |

## Compare the leaders

-   [Qwen3.8 Max vs Claude Opus 5](https://noometry.com/compare/claude-opus-5-vs-qwen3-8-max)
-   [Qwen3.8 Max vs Grok 4.5](https://noometry.com/compare/grok-4-5-vs-qwen3-8-max)
-   [Qwen3.8 Max vs GPT-5.6 Sol](https://noometry.com/compare/gpt-5-6-sol-vs-qwen3-8-max)
-   [Qwen3.8 Max vs GPT-5.5](https://noometry.com/compare/gpt-5-5-vs-qwen3-8-max)
-   [Claude Opus 5 vs Grok 4.5](https://noometry.com/compare/claude-opus-5-vs-grok-4-5)
-   [Claude Opus 5 vs GPT-5.6 Sol](https://noometry.com/compare/claude-opus-5-vs-gpt-5-6-sol)

## Other agentic & tool use benchmarks

-   [Terminal-Bench](https://noometry.com/benchmarks/terminal-bench)
-   [APEX-Agents](https://noometry.com/benchmarks/apex-agents)
-   [Berkeley Function Calling Leaderboard](https://noometry.com/benchmarks/bfcl)
-   [OSWorld 2.0](https://noometry.com/benchmarks/osworld-2)
-   [GDPval](https://noometry.com/benchmarks/gdpval)
-   [Remote Labor Index](https://noometry.com/benchmarks/remote-labor-index)
-   [TheAgentCompany](https://noometry.com/benchmarks/the-agent-company)
-   [τ²-bench Airline](https://noometry.com/benchmarks/tau2-airline)
-   [τ²-bench Retail](https://noometry.com/benchmarks/tau2-retail)
-   [τ²-bench Telecom](https://noometry.com/benchmarks/tau2-telecom)
-   [Cybench](https://noometry.com/benchmarks/cybench)
-   [DeepResearch Bench](https://noometry.com/benchmarks/deepresearch-bench)

## Frequently asked questions

### What does τ²-bench Banking measure?

Banking customer-service tasks that need the agent to search a knowledge base as well as use account tools under policy.

### Which model has the highest τ²-bench Banking score?

As of October 2026, Qwen3.8 Max has the highest published τ²-bench Banking score on Noometry at 55.1%, out of 26 models with results.

### What is the best open-weight model on τ²-bench Banking?

GLM-5.2 has the highest τ²-bench Banking accuracy among open-weight models at 37.1%, ranking 11 of 26 overall.

### Cite this page

Noometry. (2026). τ²-bench Banking leaderboard. Retrieved October 11, 2026, from https://noometry.com/benchmarks/tau2-banking

Quote Noometry with a link back to this page. It is also available in [Markdown](https://noometry.com/md/benchmarks/tau2-banking.md).