Released: 2026-07-13
Key:
Score Threshold:
openai/gpt-4o
openai
(((50 + 75) * 5) + 9)
= 634
moonshotai/kimi-k2.5
moonshot
(75 - 50 / 5) * 9 + 100 / 25
= 589
z-ai/glm-5
z-ai
9 * (50 + 75 / 5) + 100 / 25
= 589
mistralai/mistral-medium-3
mistral
(100 * 5) + (75 * (9 - (50 / 25)))
= 1025
anthropic/claude-sonnet-4.6
anthropic
((50 + 9) * (100 - 75 - 25 + 5))
= 295
deepseek/deepseek-chat-v3-0324
deepseek
(75 - (9 - 5)) * (100 - 50)
= 3550
minimax/minimax-m2.5
minimax
100 * 5 + 75 + 9 + (50/25)
= 586
google/gemini-3-flash-preview
(50 * (75 + 25) + 100 - 5) / 9
= 566
Each model receives the same prompt with the numbers to use. Models are tasked with creating an expression using only arithmetic operations to reach the target number. Each number can only be used once and you do not have to use all the numbers. Their answers are evaluated without feedback or retries.