VPCT
ReportedNo description yet: we only describe an eval set when we can do it correctly.
- Kind
- Published by a third party's leaderboard, collected by Epoch AI. The leaderboard is named and linked.
- Link
- https://cbrower.dev/vpct
- Score shown
- Epoch's column "Correct", as a percentage, higher is better; random guessing scores about 33.3%. The source gives no error bars.
- Results
- 38 in all: 0 for open-weight models, 37 for closed ones, 1 where Epoch does not say. 0 linked to a model page here.
No open-weight model has a result on this set.
| # | Model | Weights | Setting | Score | Date | Source |
|---|---|---|---|---|---|---|
| 1 | gemini-3-pro-previewGoogle DeepMind | closed | default | 91.0% | not dated | VPCT leaderboard |
| 2 | gpt-5.2-2025-12-11_xhighOpenAI | closed | effort xhigh | 84.0% | not dated | VPCT leaderboard |
| 3 | gemini-3-flash-previewGoogle DeepMind | closed | default | 72.6% | not dated | VPCT leaderboard |
| 4 | gpt-5.2-2025-12-11_highOpenAI | closed | effort high | 67.0% | not dated | VPCT leaderboard |
| 5 | gpt-5-2025-08-07_highOpenAI | closed | effort high | 66.0% | not dated | VPCT leaderboard |
| 6 | gpt-5-2025-08-07_mediumOpenAI | closed | effort medium | 63.2% | not dated | VPCT leaderboard |
| 7 | gpt-5.1-2025-11-13_highOpenAI | closed | effort high | 58.7% | not dated | VPCT leaderboard |
| 8 | o4-mini-2025-04-16_mediumOpenAI | closed | effort medium | 57.5% | not dated | VPCT leaderboard |
| 9 | gpt-5.1-2025-11-13_mediumOpenAI | closed | effort medium | 53.3% | not dated | VPCT leaderboard |
| 10 | o3-2025-04-16_mediumOpenAI | closed | effort medium | 52.0% | not dated | VPCT leaderboard |
| 11 | gemini-2.5-pro-preview-03-25Google DeepMind | closed | default | 48.0% | not dated | VPCT leaderboard |
| 12 | gemini-2.5-pro-preview-06-05Google DeepMind | closed | default | 46.4% | not dated | VPCT leaderboard |
| 13 | gemini-2.5-flash-preview-09-2025Google DeepMind | closed | default | 46.2% | not dated | VPCT leaderboard |
| 14 | gpt-4.5-preview-2025-02-27OpenAI | closed | default | 45.0% | not dated | VPCT leaderboard |
| 15 | gemini-robotics-er-1.5-preview | not stated | default | 40.8% | not dated | VPCT leaderboard |
| 16 | gemini-2.5-pro-preview-05-06Google DeepMind | closed | default | 40.5% | not dated | VPCT leaderboard |
| 17 | gpt-5-mini-2025-08-07_mediumOpenAI | closed | effort medium | 40.2% | not dated | VPCT leaderboard |
| 18 | claude-opus-4-5-20251101_32KAnthropic | closed | default | 40.0% | not dated | Tweet from VPCT creator |
| 19 | gpt-4o-2024-11-20OpenAI | closed | default | 40.0% | not dated | VPCT leaderboard |
| 20 | claude-sonnet-4-5-20250929_32KAnthropic | closed | default | 39.8% | not dated | VPCT leaderboard |
| 21 | claude-3-7-sonnet-20250219Anthropic | closed | default | 39.0% | not dated | VPCT leaderboard |
| 22 | gpt-5-mini-2025-08-07_highOpenAI | closed | effort high | 39.0% | not dated | VPCT leaderboard |
| 23 | claude-opus-4-20250514_16KAnthropic | closed | default | 38.0% | not dated | VPCT leaderboard |
| 24 | claude-sonnet-4-5-20250929Anthropic | closed | default | 38.0% | not dated | VPCT leaderboard |
| 25 | gemini-2.5-flash-preview-04-17Google DeepMind | closed | default | 38.0% | not dated | VPCT leaderboard |
| 26 | gpt-5-nano-2025-08-07_highOpenAI | closed | effort high | 37.2% | not dated | VPCT leaderboard |
| 27 | o1-2024-12-17_mediumOpenAI | closed | effort medium | 37.0% | not dated | VPCT leaderboard |
| 28 | gpt-5-nano-2025-08-07_mediumOpenAI | closed | effort medium | 35.4% | not dated | VPCT leaderboard |
| 29 | claude-3-7-sonnet-20250219_64KAnthropic | closed | default | 35.0% | not dated | VPCT leaderboard |
| 30 | claude-opus-4-1-20250805Anthropic | closed | default | 35.0% | not dated | VPCT leaderboard |
| 31 | claude-sonnet-4-20250514_32KAnthropic | closed | default | 34.0% | not dated | VPCT leaderboard |
| 32 | gpt-4o-mini-2024-07-18OpenAI | closed | default | 34.0% | not dated | VPCT leaderboard |
| 33 | claude-3-5-sonnet-20240620Anthropic | closed | default | 33.0% | not dated | VPCT leaderboard |
| 34 | claude-3-5-sonnet-20241022Anthropic | closed | default | 33.0% | not dated | VPCT leaderboard |
| 35 | claude-opus-4-1-20250805_16KAnthropic | closed | default | 33.0% | not dated | VPCT leaderboard |
| 36 | claude-opus-4-20250514Anthropic | closed | default | 33.0% | not dated | VPCT leaderboard |
| 37 | claude-sonnet-4-20250514Anthropic | closed | default | 30.0% | not dated | VPCT leaderboard |
| 38 | gemini-2.5-flash-lite-preview-09-2025Google DeepMind | closed | default | 30.0% | not dated | VPCT leaderboard |
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