gemini-3-pro-10000x-hard-high-reasoning
- Type
- dataset
- Venue
- Roman1111111
- Year
- 2026
- Source
- huggingface
- Access
- free
- Language
- en
- Added
- 2026-07-17T20:00:42.288517+00:00
- Verified
- 2026-07-17T20:00:42.288517+00:00
Summary
This dataset is a high-complexity synthetic reasoning corpus containing expert-level problems and solutions. It was generated to test the absolute limits of modern reasoning models, focusing on domains requiring multi-step logic, derivation, and synthesis of conflicting information.
Keywords
hf-dataset qa · -language-modeling · -reasoning · -code json text datasets pandas polars mlcroissant code finance legal agent chemistry art synthetic gemini-3-pro hard-reasoning mathematics physics
Topics
QA, Language Modeling, Reasoning, Code
Research notes
- downloads=84; likes=57