Released: 2026-03-18
Key:
Score Threshold:
openai/gpt-4o
openai
((75 + 4) * 8) + 100 - 25
= 707
mistralai/mistral-medium-3
mistral
((100 - 8) * (75 - (25 + 10))) + 4
= 3684
google/gemini-3-flash-preview
((75 + 8) * 10) - 100 - (25 / 4)
= 723
deepseek/deepseek-chat-v3-0324
deepseek
((100 + 10) * (8 - 4)) + 75 + 25
= 540
x-ai/grok-3-mini-beta
x-ai
((100 + 75) * 4) - (25 + 8)
= 667
z-ai/glm-5
z-ai
100 * (10 - 4) + 75 - 8
= 667
moonshotai/kimi-k2.5
moonshot
(100 - 75) * 25 + 8 * 4 + 10
= 667
google/gemini-2.0-flash-001
(75 * 10) - (100 - 8 - 4)
= 662
minimax/minimax-m2.5
minimax
100 * (10 - 4) + 75 - 8
= 667
anthropic/claude-3.7-sonnet
anthropic
(8 * 75) + (10 * 4) + 25 - 100
= 565
anthropic/claude-sonnet-4.6
anthropic
25 * (100 / 4) - 8 - 10 + 75
= 682
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.