GDPval
ReportedWork tasks from many occupations, where the model's output is compared with an expert's by expert graders.
- Kind
- Published by a third party's leaderboard, collected by Epoch AI. The leaderboard is named and linked.
- Score shown
- Epoch's column "Win Rate (%)", as a percentage, higher is better; random guessing scores about 0.0%. The source gives no error bars.
- Results
- 11 in all: 0 for open-weight models, 11 for closed ones. 0 linked to a model page here.
No open-weight model has a result on this set.
| # | Model | Weights | Setting | Score | Date | Source |
|---|---|---|---|---|---|---|
| 1 | gpt-5.2-2025-12-11_noneOpenAI | closed | effort none | 49.7% | not dated | via Epoch AI |
| 2 | claude-opus-4-5-20251101Anthropic | closed | default | 45.5% | not dated | via Epoch AI |
| 3 | claude-opus-4-1-20250805Anthropic | closed | default | 43.6% | not dated | via Epoch AI |
| 4 | claude-sonnet-4-5-20250929Anthropic | closed | default | 42.5% | not dated | via Epoch AI |
| 5 | gemini-3-pro-previewGoogle DeepMind | closed | default | 40.3% | not dated | via Epoch AI |
| 6 | gpt-5-2025-08-07_mediumOpenAI | closed | effort medium | 34.8% | not dated | via Epoch AI |
| 7 | o3-2025-04-16_mediumOpenAI | closed | effort medium | 30.8% | not dated | via Epoch AI |
| 8 | o4-mini-2025-04-16_highOpenAI | closed | effort high | 25.3% | not dated | via Epoch AI |
| 9 | gemini-2.5-proGoogle DeepMind | closed | default | 23.3% | not dated | via Epoch AI |
| 10 | grok-4-0709_highxAI | closed | effort high | 21.1% | not dated | via Epoch AI |
| 11 | gpt-4o-2024-11-20OpenAI | closed | default | 9.9% | not dated | via Epoch AI |
Open-weight rows are shaded. Dates are when the result was run or evaluated, where the data says. Data from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Epoch AI. Method