VAMPS (Analytical): leaderboard

Metric: Accuracy (%) on the English versions of the 218 Konkour-seed questions (Iranian university entrance exam algebra and calculus, chosen so that plotting reveals the answer; four-option multiple choice), direct analytical solving without tools (the standard protocol); thinking disabled for every model, temperature 0, answers read from a required JSON block; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 18 models tracked.

Top models

#ModelScore
1Claude Sonnet 4.698.2
2Claude Opus 4.798.2
3Qwen 3.5 397B A17B (Non-reasoning)97.2
4Gemma 4 31B (Non-reasoning)97.2
5Qwen 3.5 27B (Non-reasoning)95.9
6Qwen 3 VL 8B Instruct92.2
7Qwen 3 VL 32B Instruct92.2
8Gemini 2.5 Flash (Non-reasoning)90.4
9Gemma 4 26B A4B (Non-reasoning)90.4
10GPT-5.4 (Non-reasoning)89
11Gemma 3 27B77.5
12Ministral 3 14B75.7
13Gemma 3 12B68.3
14Ministral 3 8B67.4
15Qwen 3.5 35B A3B (Non-reasoning)64.7

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