Released: 2026-03-03
Key:
Score Threshold:
mistralai/mistral-medium-3
mistral
((75 + 50) * (7 + 5)) + (25 * 6)
= 1650
google/gemini-2.0-flash-001
(75 + 25) * (7 + 6)
= 1300
openai/gpt-4o
openai
(75 + 50 + 25) * 6 + 7 + 5
= 912
deepseek/deepseek-chat-v3-0324
deepseek
(75 + 25) * (7 + 6 - 5)
= 800
anthropic/claude-3.7-sonnet
anthropic
(75 + 6) * 50 / 5
= 810
google/gemini-3-flash-preview
(75 * (7 + 5)) - 50
= 850
x-ai/grok-3-mini-beta
x-ai
75 * (6 + 5) + 25
= 850
minimax/minimax-m2.5
minimax
(75 * 7) + (50 * 6) + 25
= 850
moonshotai/kimi-k2.5
moonshot
(75 + 25) * (7 + 6 - 5) + 50
= 850
z-ai/glm-5
z-ai
(25 + 5 - 7 - 6) * 50
= 850
anthropic/claude-sonnet-4.6
anthropic
((75 + 50) * 6) + 25 * (7 - 5)
= 800
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.