Agentic & Tool Use benchmark
GDPval leaderboard
As of October 2026, GPT-5.2 has the highest published GDPval score on Noometry at 49.7%, out of 11 models with results.
Last verified
About GDPval
Real work products from 44 occupations, graded by industry experts against the work of human professionals. Scores are win-or-tie rates.
- Category
- Agentic & Tool Use
- Introduced
- 2025
- Format
- Expert-graded deliverables
- Unit
- Percent (random guessing ≈ 0%)
- Official site
- openai.com
Top 11 models
- GPT-5.2 49.7%
- Claude Opus 4.5 45.5%
- Claude Opus 4.1 43.6%
- Claude Sonnet 4.5 42.5%
- Gemini 3 Pro 40.3%
- GPT-5 34.8%
- o3 30.8%
- o4-mini 25.3%
- Gemini 2.5 Pro 23.3%
- Grok 4 21.1%
- GPT-4o 9.9%
Sponsored placements are available on pages like this one. Advertise on Noometry
All results
| # | Model | Provider | Score | Setting | Source | Date |
|---|---|---|---|---|---|---|
| 1 | GPT-5.2 | OpenAI | 49.7% | none | Epoch AI | |
| 2 | Claude Opus 4.5 | Anthropic | 45.5% | Epoch AI | ||
| 3 | Claude Opus 4.1 | Anthropic | 43.6% | Epoch AI | ||
| 4 | Claude Sonnet 4.5 | Anthropic | 42.5% | Epoch AI | ||
| 5 | Gemini 3 Pro |  Google | 40.3% | Epoch AI | ||
| 6 | GPT-5 | OpenAI | 34.8% | medium | Epoch AI | |
| 7 | o3 | OpenAI | 30.8% | medium | Epoch AI | |
| 8 | o4-mini | OpenAI | 25.3% | high | Epoch AI | |
| 9 | Gemini 2.5 Pro | Google | 23.3% | Epoch AI | ||
| 10 | Grok 4 | xAI | 21.1% | high | Epoch AI | |
| 11 | GPT-4o | OpenAI | 9.9% | Epoch AI |
Compare the leaders
Frequently asked questions
What does GDPval measure?
Real work products from 44 occupations, graded by industry experts against the work of human professionals. Scores are win-or-tie rates.
Which model has the highest GDPval score?
As of October 2026, GPT-5.2 has the highest published GDPval score on Noometry at 49.7%, out of 11 models with results.