SciR Benchmark

What is SciR?

SciR evaluates LLMs on deduction, induction, and causal abduction in scientific settings, with tasks generated from formal objects and rendered into…

Released

2026-06-11

Evaluates

General AI, Reasoning, Vision & 3D

Openness

unknown

Importer

Claire Radar

Review status

unreviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2606.13020

SciR paper

• Paper https://arxiv.org/abs/2606.13020

Which sources cite SciR?

• SciR

Related General AI, Reasoning, Vision & 3D benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 3DCodeBench
• 5-Dialects-BN
• A-OKVQA
• A3S-Bench
• AARRI-Bench
• ActBench
• ActiveFly-Bench