AI vs Puzzles

Back to Home
Puzzle#147

Countdown

Released: 2026-03-23


11
Models Tested
3
Solved
9
LLM Providers

Key:

Used
Unused

Score Threshold:

Within 7

openai/gpt-4o

openai

680Target
75
50
100
25
10
5

(75 * 10) - 100 + 5

= 655

Failed

mistralai/mistral-medium-3

mistral

680Target
75
50
100
25
10
5

((100 + 50) * (75 - 25)) - (10 * 5)

= 7450

Failed

anthropic/claude-sonnet-4.6

anthropic

680Target
75
50
100
25
10
5

75 * (100 - 50 - 25) / 5 + 10

= 385

Failed

deepseek/deepseek-chat-v3-0324

deepseek

680Target
75
50
100
25
10
5

(100 + 25) * 5 + 75 - 10

= 690

Failed

Minimax

minimax/minimax-m2.5

minimax

680Target
75
50
100
25
10
5

75 * 10 - (50 + 25 - 5)

= 680

Perfect Solution!

MoonshotAI

moonshotai/kimi-k2.5

moonshot

680Target
75
50
100
25
10
5

75*10 - 50 - 25 + 5

= 680

Perfect Solution!

x-ai/grok-3-mini-beta

x-ai

680Target
75
50
100
25
10
5

((100 * (50 / 5)) - (75 + (25 * 10)))

= 675

Close; 5 away

Z.ai

z-ai/glm-5

z-ai

680Target
75
50
100
25
10
5

(75 * 10) - 50 - 25 + 5

= 680

Perfect Solution!

google/gemini-3-flash-preview

google

680Target
75
50
100
25
10
5

75 + 10 * (50 + 100 - 25) / 5

= 325

Failed

google/gemini-2.0-flash-001

google

680Target
75
50
100
25
10
5

(100 - 10) * (5 + 75 / 25)

= 720

Failed

anthropic/claude-3.7-sonnet

anthropic

680Target
75
50
100
25
10
5

(75 + 100 + 5) * 50 / 25

= 360

Failed

Methodology Note

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.

Leaderboard