AI vs Puzzles

Back to Home
Puzzle#125

Countdown

Released: 2026-03-01


11
Models Tested
1
Solved
9
LLM Providers

Key:

Used
Unused

Score Threshold:

Within 7

deepseek/deepseek-chat-v3-0324

deepseek

674Target
75
100
50
25
4
2

(100 + 75) * 4 + 50 + 25 - 2

= 773

Failed

openai/gpt-4o

openai

674Target
75
100
50
25
4
2

(100 + 75 + 50) * 4 - 2

= 898

Failed

mistralai/mistral-medium-3

mistral

674Target
75
100
50
25
4
2

(100 * 75) - (50 * 25) - (4 * 2)

= 6242

Failed

google/gemini-3-flash-preview

google

674Target
75
100
50
25
4
2

(75 * (100 - 50) / 25) - 2 + 4

= 152

Failed

anthropic/claude-3.7-sonnet

anthropic

674Target
75
100
50
25
4
2

75 * 100 / 25 * 2 + 50 - 4

= 646

Failed

google/gemini-2.0-flash-001

google

674Target
75
100
50
25
4
2

(75 + 100) * 4 - 25 + 50 - 2

= 723

Failed

Minimax

minimax/minimax-m2.5

minimax

674Target
75
100
50
25
4
2

100 * (4 + 2) + 75

= 675

Close; 1 away

x-ai/grok-3-mini-beta

x-ai

674Target
75
100
50
25
4
2

100 * (4 + 2) + 75

= 675

Close; 1 away

MoonshotAI

moonshotai/kimi-k2.5

moonshot

674Target
75
100
50
25
4
2

(100 + 4) * (75 / 25 * 2) + 50

= 674

Perfect Solution!

anthropic/claude-sonnet-4.6

anthropic

674Target
75
100
50
25
4
2

75 * (4 + 100 / 50) + 25 - 2

= 473

Failed

Z.ai

z-ai/glm-5

z-ai

674Target
75
100
50
25
4
2

(100 + 75) * 4 - 25 - 2

= 673

Close; 1 away

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