DSGym - DSPredict Hard - Any Medal: leaderboard
DSPredict-Hard, DSGym's split of recent, still-open Kaggle competitions with complex data and objectives, scored on the private leaderboard and run by DSGym's default CodeAct agent (Nie et al., 2026) in an isolated Jupyter container at temperature 0 with all tools disabled. Share of competitions whose submission earns any medal, in percent; higher is better.
Metric: Any Medal Rate (%; private leaderboard). Source: arxiv.org. Saturation forecast: Around 2030. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.1 (High) | 4.8 | #225 (GPT-5.1) |
| 2 | GPT-5.1 (Medium) | 4.8 | #225 (GPT-5.1) |
| 3 | Claude Sonnet 4 | 2.4 | #397 |
| 4 | DeepSeek V3.1 | 2.4 | #500 |
| 5 | GPT-5.1 (Non-reasoning) | 2.4 | #225 (GPT-5.1) |
| 6 | Claude Sonnet 4.5 | 0 | #244 |
| 7 | Qwen 2.5 7B Instruct | 0 | #2583 |
| 8 | Kimi K2 0905 | 0 | #555 |
| 9 | GPT-5 (Medium) | 0 | #163 (GPT-5) |
| 10 | Qwen 3 4B Instruct | 0 | #1660 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=dsgym-dspredict-hard-any-medal · How It Works · Data refreshed daily, snapshot 2026-11-12.