# AA-AnalystAgent Contamination Audit

> AA-AnalystAgent scores 8.3/10 (Tier A, Low Risk) on BenchAudit as of November 2026, reflecting a strictly private held-out test set and self-reported or live-network evaluation.

- **Canonical URL:** https://benchaudit.org/benchmarks/aa-analystagent/
- **Benchmark ID:** `aa-analystagent`
- **Hub:** Artificial Analysis
- **Category:** Knowledge Work
- **Domains:** Finance, Business, Math
- **Skills:** Agentic, Tool Use, Knowledge Work, Coding, Reasoning
- **BenchAudit Score:** 8.3/10 (Tier A - Low Risk)
- **Last Audited:** 2026-11-17 (Ref 2026-11)

## Factor Scores

| Factor | Score (/10) | Weight |
| --- | ---: | ---: |
| Privacy | 10 | 40% |
| Originality | 7.2 | 25% |
| Freshness | 10 | 20% |
| Grading | 3.3 | 15% |

## Primary-Source Evidence

### Privacy
> "AA-AnalystAgent is a privately held benchmark; the question set, reference answers, and source files are not publicly released, to limit contamination risk"
- Source: https://artificialanalysis.ai/methodology/intelligence-benchmarking

### Originality
> "80 quantitative questions across 14 business and scientific domains, including environmental reporting, trade and commodity statistics, healthcare expenditure reports, hydrology and weather data, government appropriations, energy cost models, financial models, and project schedules"
- Source: https://artificialanalysis.ai/methodology/intelligence-benchmarking

### Freshness
> "August 10, 2026 Announcing AA-AnalystAgent: an agentic benchmark for quantitative analysis on real-world spreadsheets and documents"
- Source: https://artificialanalysis.ai/articles/aa-analyst-agent

### Grading
> "The agent is provided a small toolset covering code execution in an isolated Linux sandbox (Python 3.12 with the question's reference files mounted and a pinned set of standard Python data analysis libraries preinstalled), URL fetching, image viewing for vision-capable models, and final-answer submission."
- Source: https://artificialanalysis.ai/methodology/intelligence-benchmarking

## Top Frontier Models on AA-AnalystAgent

| Rank | Model | Creator | Percentile |
| ---: | --- | --- | ---: |
| #1 | [GPT-6.2 Astra](https://benchaudit.org/models/gpt-6-2-astra__max/) | OpenAI | 97.2p |
| #2 | [Gemini 3.7 Flash](https://benchaudit.org/models/gemini-3-7-flash__high/) | Google | 93.7p |
| #3 | [Claude Fable 5.1](https://benchaudit.org/models/claude-fable-5-1__max/) | Anthropic | 88.3p |
| #3 | [Claude Sonnet 5.5](https://benchaudit.org/models/claude-sonnet-5-5__max/) | Anthropic | 88.3p |
| #5 | [Claude Opus 5.5](https://benchaudit.org/models/claude-opus-5-5__max/) | Anthropic | 83.0p |
| #6 | [Claude Haiku 5.5](https://benchaudit.org/models/claude-haiku-5-5__max/) | Anthropic | 77.7p |
| #6 | [GPT-6.2 Sol](https://benchaudit.org/models/gpt-6-2-sol__max/) | OpenAI | 77.7p |
| #8 | [Grok 4.7](https://benchaudit.org/models/grok-4-7__xhigh/) | SpaceXAI | 72.4p |
| #9 | [GPT-6 Astra](https://benchaudit.org/models/gpt-6-astra__max/) | OpenAI | 67.0p |
| #9 | [Inkling 2 Small](https://benchaudit.org/models/inkling-2-small/) | Thinking Machines | 67.0p |
