GDP.pdf
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://surgehq.ai/benchmarks/gdp-pdf
- Score shown
- Epoch's column "GDP.pdf score", as a percentage, higher is better. The source gives no error bars.
- Results
- 19 in all: 6 for open-weight models, 13 for closed ones. 5 linked to a model page here.
Best open-weight result: moonshotai/Kimi-K3.1-Pro, 46.6%, number 12 of 19 results.
| # | Model | Weights | Setting | Score | Date | Source |
|---|---|---|---|---|---|---|
| 1 | gemini-4-carbon_highGoogle DeepMind | closed | effort high | 56.4% | not dated | surgehq.ai |
| 2 | muse-wonder-1.0_maxMeta AI | closed | effort max | 54.8% | not dated | surgehq.ai |
| 3 | gemini-4-argon_highGoogle DeepMind | closed | effort high | 53.6% | not dated | surgehq.ai |
| 4 | muse-spark-1.3_maxMeta AI | closed | effort max | 53.0% | not dated | surgehq.ai |
| 5 | gpt-6.1-sol_maxOpenAI | closed | effort max | 51.6% | not dated | surgehq.ai |
| 6 | claude-sonnet-5-5_maxAnthropic | closed | effort max | 51.2% | not dated | surgehq.ai |
| 7 | gpt-6-astra_maxOpenAI | closed | effort max | 50.4% | not dated | surgehq.ai |
| 8 | claude-fable-5-1_maxAnthropic | closed | effort max | 49.4% | not dated | surgehq.ai |
| 9 | muse-ember-1.0_maxMeta AI | closed | effort max | 49.2% | not dated | surgehq.ai |
| 10 | claude-opus-5-5_maxAnthropic | closed | effort max | 48.8% | not dated | surgehq.ai |
| 11 | grok-4.7_xhighxAI | closed | effort xhigh | 48.8% | not dated | surgehq.ai |
| 12 | moonshotai/Kimi-K3.1-ProEpoch: kimi-k3.1-pro_max, Moonshot | open | effort max | 46.6% | not dated | surgehq.ai |
| 13 | gemini-3.8-flash_highGoogle DeepMind | closed | effort high | 44.2% | not dated | surgehq.ai |
| 14 | thinkingmachines/Inkling-2-SmallEpoch: inkling-2-small_xhigh, Thinking Machines | open | effort xhigh | 43.6% | not dated | surgehq.ai |
| 15 | qwen3.8-max_xhighAlibaba | closed | effort xhigh | 42.8% | not dated | surgehq.ai |
| 16 | moonshotai/Kimi-K3Epoch: kimi-k3_max, Moonshot | open | effort max | 39.8% | not dated | surgehq.ai |
| 17 | zai-org/GLM-5.3Epoch: glm-5.3_max, Z.ai (Zhipu AI) | open | effort max | 36.6% | not dated | surgehq.ai |
| 18 | tencent/Hy4-previewEpoch: hy4-preview_unknown, Tencent | open | effort unknown | 35.0% | not dated | surgehq.ai |
| 19 | mistral-large-4-0_highMistral AIno page match | open | effort high | 32.8% | not dated | surgehq.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