Released: 2026-07-14
Key:
Score Threshold:
google/gemini-3-flash-preview
(8 * 75) - (4 - (50 / 25))
= 598
openai/gpt-4o
openai
((75 * 8) - 4) - 25
= 571
mistralai/mistral-medium-3
mistral
((75 + 50) * (8 + 4)) - (100 / 25)
= 1496
deepseek/deepseek-chat-v3-0324
deepseek
(100 * (75 - (50 / 25))) - (8 + 4)
= 7288
minimax/minimax-m2.5
minimax
100 + 50 + 25 * 8 + 4 * 75
= 650
moonshotai/kimi-k2.5
moonshot
50 * (8 + 4) - 100 / (75 + 25)
= 599
z-ai/glm-5
z-ai
(8 + 4) * 50 - (75 + 25) / 100
= 599
anthropic/claude-sonnet-4.6
anthropic
75 * 8 - 4 + 25 - 50 + 100
= 671
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.