AI vs Puzzles

Back to Home
Puzzle#1707

Wordle

Released: 2026-02-20

7
Models Tested
3
Solved
5.43
Avg. Attempts
7
LLM Providers
Correct
Present
Absent

openai/gpt-4o

openai

Failed

c
r
a
n
e
s
w
a
m
p
s
a
n
d
y
s
n
a
i
l
s
n
e
a
k
s
n
a
c
k

mistralai/mistral-medium-3

mistral

Failed

c
r
a
n
e
p
l
a
n
t
b
a
n
t
y
s
a
i
n
t
s
a
t
a
n
s
a
n
t
a

meta-llama/llama-3.3-70b-instruct

meta

Failed

h
o
u
s
e
b
l
a
s
t
t
o
a
s
t
t
a
l
e
s
s
t
a
i
n
s
t
a
t
s

google/gemini-2.0-flash-001

google

Failed

c
r
a
n
e
b
l
a
n
d
o
a
k
e
n
k
a
n
j
i
a
n
k
l
e
k
n
a
c
k

deepseek/deepseek-chat-v3-0324

deepseek

4/6

c
r
a
n
e
p
l
a
n
t
s
l
a
n
t
s
t
a
n
k

anthropic/claude-3.7-sonnet

anthropic

4/6

c
r
a
t
e
p
l
a
n
t
b
a
n
k
s
s
t
a
n
k

x-ai/grok-3-mini-beta

x-ai

6/6

c
r
a
n
e
s
l
a
n
t
s
t
a
n
d
s
t
a
n
t
s
t
a
r
k
s
t
a
n
k

Methodology Note

Each model receives the same prompt structure with Wordle rules and feedback format. Models are given a maximum of 6 attempts with standard positional feedback after each guess to help refine their next attempt.

Leaderboard