Released: 2026-04-22
Key:
Score Threshold:
openai/gpt-4o
openai
((50 * 6) * (4 - 2)) - 25/4
= 593
google/gemini-2.0-flash-001
(50 * (6 + 4)) - (25 - 4) * 2
= 458
mistralai/mistral-medium-3
mistral
((25 * (50 - (6 / 2))) + (4 * 4))
= 1191
anthropic/claude-3.7-sonnet
anthropic
50 * 25 - 4 - (6 / 2)
= 1243
deepseek/deepseek-chat-v3-0324
deepseek
((50 + 6) * (25 - 4)) - 4
= 1172
x-ai/grok-3-mini-beta
x-ai
50 * (6 + 4)
= 500
z-ai/glm-5
z-ai
(50 - 25) * (6 * 4 - 4) - 2
= 498
minimax/minimax-m2.5
minimax
50 * (6 + 4) + 25 - 2 - 4
= 519
moonshotai/kimi-k2.5
moonshot
(50 + 2) * (6 + 4) - (25 - 4)
= 499
anthropic/claude-sonnet-4.6
anthropic
50 * (4 + 6) - 4 + 25 - 2
= 519
google/gemini-3-flash-preview
(25 * 50 / 2) - (6 * 4) + 4
= 605
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.