Released: 2026-03-12
Key:
Score Threshold:
openai/gpt-4o
openai
(9 * 50) + (10 * (9 - 1))
= 530
deepseek/deepseek-chat-v3-0324
deepseek
(100 + 9) * (10 - (9 - 1)) + 50
= 268
mistralai/mistral-medium-3
mistral
((100 + 9) * (9 + 50)) - 10
= 6421
x-ai/grok-3-mini-beta
x-ai
(50 * 9) - (10 - 9) * 1
= 449
z-ai/glm-5
z-ai
(100 - 9) * (50 / 10) - 9
= 446
google/gemini-2.0-flash-001
(100 - 50) * 9 - 10 + 9 + 1
= 450
anthropic/claude-3.7-sonnet
anthropic
(9 * 50) - 9 + 10 + 100 + 1
= 552
moonshotai/kimi-k2.5
moonshot
(100 - 9) * (50 / 10) - 9
= 446
minimax/minimax-m2.5
minimax
(100 + 50) * 9 + 10 - 9 - 1
= 1350
anthropic/claude-sonnet-4.6
anthropic
100 * (1 + 9/9) + 50 - 10
= 240
google/gemini-3-flash-preview
(50 * 9) - ((100 / 10) / (9 - 1))
= 448
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.