Released: 2026-05-07
Key:
Score Threshold:
google/gemini-2.0-flash-001
(100 - 10) * (3 + 4)
= 630
openai/gpt-4o
openai
((100 * 4) + (10 * 3)) + 3
= 433
google/gemini-3-flash-preview
(100 + 4) * (10 + 3 - 9)
= 416
anthropic/claude-3.7-sonnet
anthropic
(9 * (100 / 10)) * (4 + 3)
= 630
deepseek/deepseek-chat-v3-0324
deepseek
((100 + 10) * (4 - (3 / 3))) * 9
= 2970
anthropic/claude-sonnet-4.6
anthropic
100 * 4 + 9 * (10 - 3 - 3)
= 436
x-ai/grok-3-mini-beta
x-ai
3 * (100 + 10 * 4)
= 420
minimax/minimax-m2.5
minimax
(100 + 4 * 10) * 3
= 420
moonshotai/kimi-k2.5
moonshot
(100 + 4 * 10) * (9 - 3 - 3)
= 420
mistralai/mistral-medium-3
mistral
(100 * (10 - 3)) * (4 - (9 / 3))
= 700
z-ai/glm-5
z-ai
(9 * 4 + 3 + 3) * 10
= 420
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.