SciR - Deduction Easy (Obfuscated): leaderboard

Metric: Accuracy (%) on the 200 easy-tier SciR deduction tasks (deduction over developmental-biology lineage syllogisms (3-way answer)), rendered as obfuscated multi-document scientific discourse (Obf: lab notes, drug labels, records and tables); the model reasons with chain-of-thought directly from the task text, temperature 0; answers are checked against the generating formal object; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 6 models tracked.

Top models

#ModelScore
1DeepSeek R162.5
2Qwen 3 30B A3B 2507 Instruct47
3GPT-4o44
4Olmo 3.1 32B Instruct33
5Llama 3.3 70B Instruct28

Interactive version: theaggregate.ai/benchmark?slug=scir-deduction-easy-obfuscated · How It Works · Data refreshed daily, snapshot 2026-09-29.