SeekerGym - ML Surveys: leaderboard

SeekerGym (2026) measures how completely an information-seeking agent recovers the information goals of a hidden source document through search queries over its passages. This board is completeness in percent over 80 machine-learning survey papers, counting deduplicated retrieved passages, over three seeds; higher is better.

Metric: Information-gathering completeness (%): share of a survey paper's information goals recovered by the agent's deduplicated query trajectory, 80 machine-learning survey papers, 3 seeds; higher is better. Source: arxiv.org. Saturation forecast: Around 2034. 6 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Gemini 2.0 Flash29.2#723
2Nemotron 3 Nano28.9#2025
3GPT-4o Mini24.5#1667
4GPT-OSS-20B24.1#1183
5Qwen 3 235B A22B 2507 (Thinking)23.1#466
6GPT-OSS-120B22.2#704

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=seekergym-ml-surveys · How It Works · Data refreshed daily, snapshot 2026-11-21.