MERRIN (No Search): leaderboard

MERRIN (2026) asks 162 human-written questions whose single short answers depend on non-text web evidence (images, video, audio or tables) found among noisy and conflicting sources, with no hint of the modality in the question. Closed-book setting with no search tool: accuracy over three runs, judged against the gold answer, in percent; higher is better.

Metric: Accuracy (%) with no search or browsing tool, on MERRIN's 162 human-annotated questions with unique short answers that need non-text web evidence (image, video, audio or table) and multi-hop reasoning or conflict resolution across noisy sources; accuracy against the gold answer by an LLM judge with the BrowseComp prompt, mean of three runs; higher is better. Source: arxiv.org. Saturation forecast: Around 2028. 9 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Gemini 3.1 Pro (Preview)24.7#102
2Gemini 3 Pro23.5#146
3Gemini 3 Flash19.1#180
4GPT-5.4 Mini14#409
5Gemini 3.1 Flash Lite12.8#377
6Qwen 3 235B A22B12.1#633
7Qwen 3 4B10.3#2379
8GPT-5.4 Nano9.9#653
9Qwen 3 30B A3B8#1165

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=merrin-no-search · How It Works · Data refreshed daily, snapshot 2026-11-22.