Released: 2026-03-20
Key:
Score Threshold:
deepseek/deepseek-chat-v3-0324
deepseek
(50 * 10) + (100 - 4)
= 596
x-ai/grok-3-mini-beta
x-ai
(((75 - 25) * 10) + (4 + (100 / 50)))
= 506
minimax/minimax-m2.5
minimax
50*10+4+100/(75-25)
= 506
z-ai/glm-5
z-ai
(75 - 25) * 10 + 4 + 100 / 50
= 506
openai/gpt-4o
openai
(100 * 4) + 50 + 10 + (75 / 25)
= 463
mistralai/mistral-medium-3
mistral
((100 + 75) * 4) + (50 / 10) + 25
= 730
anthropic/claude-3.7-sonnet
anthropic
(50 * 10) + (75 - 25) * 4
= 700
moonshotai/kimi-k2.5
moonshot
(75 + 50) * 4 + 10 - (100 / 25)
= 506
google/gemini-2.0-flash-001
(75 - 25) * 10 + 4 + 100 + 50
= 654
google/gemini-3-flash-preview
50 + (100 * 4) + 75 - (25 / 10)
= 522
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.