SPUR - Trend Analysis: leaderboard

SPUR (2026) tests multimodal models with multiple-choice questions on multi-panel biomedical experimental figures from PubMed Central. Accuracy in percent on trend analysis questions, which ask the model to interpret changes across panels of the same type; higher is better.

Metric: Accuracy (%) on SPUR the 1,357 trend analysis multiple-choice questions about multi-panel biomedical experimental figures from PubMed Central papers (questions drafted by GPT-4o, kept only when GPT-4o failed them in at least six of ten text-only attempts, then expert-reviewed): the questions ask the model to interpret directional changes across panels of the same type; higher is better. Source: arxiv.org. Saturation forecast: Not forecast. 19 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Ministral-3-14B-Instruct-251257.79#1651
2Ministral-3-8B-Instruct-251257.49#1923
3GLM-4.5V55.71#756
4Qwen 3 VL 30B A3B (Thinking)54.85#727
5Gemini 2.5 Pro (Preview 06-05)53.3#222
6Llama 4 Maverick51.88#1113
7Qwen 2.5 VL 72B Instruct51.87#849
8Mistral Small 3.151.82#1734
9Claude 3.7 SonnetThinking51.3#455
10GPT-5.151.18#248
11Gemini 3 Pro (Preview)51.04#123
12Qwen 3 VL 30B A3B Instruct50.41#903
13InternVL3-78B49.52#765
14Gemma 3 27B (IT)48.59#1341
15O4 MiniHigh48.37#372

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

Interactive version: theaggregate.ai/benchmark?slug=spur-trend-analysis · How It Works · Data refreshed daily, snapshot 2026-11-22.