AI vs Puzzles

Back to Home
Puzzle#158

Countdown

Released: 2026-04-03


11
Models Tested
4
Solved
9
LLM Providers

Key:

Used
Unused

Score Threshold:

Within 7

google/gemini-2.0-flash-001

google

354Target
50
25
1
3
9
3

((50 * 3) + 25) + 9

= 184

Failed

openai/gpt-4o

openai

354Target
50
25
1
3
9
3

((25 + 3) * 9) + 3 - 50

= 205

Failed

google/gemini-3-flash-preview

google

354Target
50
25
1
3
9
3

(50 + 9) * (3 + 3)

= 354

Perfect Solution!

mistralai/mistral-medium-3

mistral

354Target
50
25
1
3
9
3

((50 + 25) * (9 + 3)) * (3 - 1)

= 1800

Failed

anthropic/claude-3.7-sonnet

anthropic

354Target
50
25
1
3
9
3

(50 * 9) - (25 - 3 - 1)

= 429

Failed

deepseek/deepseek-chat-v3-0324

deepseek

354Target
50
25
1
3
9
3

(50 * (9 - (1 / 25))) + 3

= 451

Failed

Z.ai

z-ai/glm-5

z-ai

354Target
50
25
1
3
9
3

(3 + 3) * (50 + 9)

= 354

Perfect Solution!

MoonshotAI

moonshotai/kimi-k2.5

moonshot

354Target
50
25
1
3
9
3

(50 + 9) * (3 + 3)

= 354

Perfect Solution!

x-ai/grok-3-mini-beta

x-ai

354Target
50
25
1
3
9
3

50 * (9 - 3 + 1) + 3

= 353

Close; 1 away

Minimax

minimax/minimax-m2.5

minimax

354Target
50
25
1
3
9
3

(50 + 9) * (3 + 3)

= 354

Perfect Solution!

anthropic/claude-sonnet-4.6

anthropic

354Target
50
25
1
3
9
3

50 * (9 - 3) + 25 * (3 - 1)

= 350

Close; 4 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