# GDP.pdf Contamination Audit

> GDP.pdf scores 3.6/10 (Tier C, High Risk) on BenchAudit as of November 2026, reflecting publicly downloadable plaintext test files and standardized independent third-party grading.

- **Canonical URL:** https://benchaudit.org/benchmarks/gdp-pdf/
- **Benchmark ID:** `gdp-pdf`
- **Hub:** Artificial Analysis / Surge AI
- **Category:** Knowledge Work
- **Domains:** Business, Finance, Legal, Medical, Science
- **Skills:** Knowledge Work, Multimodal, Long Context, Reasoning, Writing
- **BenchAudit Score:** 3.6/10 (Tier C - High Risk)
- **Last Audited:** 2026-11-17 (Ref 2026-11)

## Factor Scores

| Factor | Score (/10) | Weight |
| --- | ---: | ---: |
| Privacy | 0 | 40% |
| Originality | 7.2 | 25% |
| Freshness | 3.5 | 20% |
| Grading | 7.3 | 15% |

## Primary-Source Evidence

### Privacy
> "All 100 examples on Hugging Face, released under Apache 2.0; third-party PDFs retain their original rights"
- Source: https://arxiv.org/html/2607.11192v1

### Originality
> "It consists of question-document pairs authored by working professionals in ten fields, and a candidate question was kept only when at least two frontier multimodal models failed it in a way that mattered: a wrong answer, missed decisive evidence, or a fabricated claim, rather than a superficial difference such as style."
- Source: https://arxiv.org/abs/2607.11192

### Freshness
> "All numbers are a frozen snapshot taken at the benchmark's public release on April 14, 2026."
- Source: https://arxiv.org/html/2607.11192v1

### Grading
> "Every model receives the complete extracted text of every page. When supported by the endpoint, we send page images first in page order, followed by the task and complete extracted text in a single user message."
- Source: https://artificialanalysis.ai/methodology/intelligence-benchmarking

## Top Frontier Models on GDP.pdf

| Rank | Model | Creator | Percentile |
| ---: | --- | --- | ---: |
| #1 | [Muse Wonder Full](https://benchaudit.org/models/muse-wonder-full__max/) | Meta | 100.0p |
| #2 | [Kimi K3.1 Pro](https://benchaudit.org/models/kimi-k3-1-pro__max/) | Kimi | 98.0p |
| #3 | [GPT-6 Astra](https://benchaudit.org/models/gpt-6-astra__max/) | OpenAI | 94.9p |
| #3 | [GPT-6.1 Sol](https://benchaudit.org/models/gpt-6-1-sol__max/) | OpenAI | 94.9p |
| #5 | [Muse Wonder](https://benchaudit.org/models/muse-wonder__max/) | Meta | 91.8p |
| #6 | [GPT-6.2 Sol](https://benchaudit.org/models/gpt-6-2-sol__max/) | OpenAI | 89.8p |
| #7 | [GPT-6.2 Astra](https://benchaudit.org/models/gpt-6-2-astra__max/) | OpenAI | 87.8p |
| #8 | [GPT-6.1 Terra](https://benchaudit.org/models/gpt-6-1-terra__max/) | OpenAI | 85.7p |
| #9 | [Gemini 4 Carbon](https://benchaudit.org/models/gemini-4-carbon__high/) | Google | 83.7p |
| #10 | [Gemini 4 Neon](https://benchaudit.org/models/gemini-4-neon__xhigh/) | Google | 81.6p |
