Released: 2026-04-11
Key:
Score Threshold:
anthropic/claude-3.7-sonnet
anthropic
4 * (75 - 5 + 50 - 25 - 100)
= -20
moonshotai/kimi-k2.5
moonshot
(75 - 25) * 5 + 100 / 50
= 252
z-ai/glm-5
z-ai
(75 - 25) * 5 + 100 / 50
= 252
minimax/minimax-m2.5
minimax
75 * 4 - 100 + 50 - 25 - 5
= 220
openai/gpt-4o
openai
((75 + 25) * 4) - 50 + 5
= 355
google/gemini-2.0-flash-001
(75 + 25) * (5 - 4)
= 100
deepseek/deepseek-chat-v3-0324
deepseek
((100 - 25) * 4) - (75 - 50 - 5)
= 280
x-ai/grok-3-mini-beta
x-ai
50 * 5
= 250
anthropic/claude-sonnet-4.6
anthropic
75 * 4 - 50 + 25 + 5 - 100
= 180
mistralai/mistral-medium-3
mistral
(100 + 75 + 50 + 25) * (5 - 4)
= 250
google/gemini-3-flash-preview
100 + (75 * (50 + 25)) / (5 * 4)
= 381
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.