SciConvBench - Inconsistency Resolution - PDEs - Final Resolution: leaderboard
Metric: Final Resolution Rate (%): share of pdes inconsistency resolution cases (requests with planted conflicting information) whose final specification resolves every annotated issue, the assistant may ask one clarification question per turn of a simulated scientist (Claude Sonnet 4.6, answering only from the hidden reference specification) for at most 11 turns before producing the final task specification, guided scientist-mode system prompt, judged by Gemini 2.5 Pro against the annotated issues; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 93.2 |
| 2 | Gemini 2.5 Flash | 89 |
| 3 | GPT-OSS-120B | 72.6 |
| 4 | GPT-5.2 | 68.5 |
| 5 | Claude Sonnet 4.6 | 31.5 |
Interactive version: theaggregate.ai/benchmark?slug=sciconvbench-inconsistency-resolution-pdes-final-resolution · How It Works · Data refreshed daily, snapshot 2026-10-07.