PDEAgent-Bench - Navier-Stokes: leaderboard
Metric: Case pass rate (%; 28 DOLFINx cases, single-shot, temperature 0). Source: arxiv.org. Saturation forecast: Around December 2026. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 | 57.1 |
| 2 | GPT-5.4 | 50 |
| 3 | Gemini 3.1 Pro (Preview) | 21.4 |
| 4 | Qwen 3.6 Plus | 17.9 |
| 5 | DeepSeek V3.2 | 0 |
Interactive version: theaggregate.ai/benchmark?slug=pdeagent-bench-navier-stokes · How It Works · Data refreshed daily, snapshot 2026-09-25.