AI vs Puzzles

Back to Home
Puzzle#1687

Wordle

Released: 2026-01-31

8
Models Tested
1
Solved
5.88
Avg. Attempts
8
LLM Providers
Correct
Present
Absent

openai/gpt-4o

openai

Failed

c
r
a
n
e
f
l
a
i
r
l
l
a
m
a
l
e
a
k
y
b
a
l
s
a
g
u
l
c
h

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

meta

Failed

h
o
u
s
e
b
o
a
s
t
c
o
a
l
t
v
a
u
l
t
f
o
l
k
t
m
o
l
t
z

google/gemini-2.0-flash-001

google

Failed

c
r
a
n
e
b
l
a
s
t
f
l
a
i
t
c
l
a
p
t
h
a
l
f
t
u
l
t
r
a

deepseek/deepseek-chat-v3-0324

deepseek

Failed

c
r
a
n
e
s
t
a
m
p
a
d
a
p
t
a
l
e
r
t
a
l
p
h
a
a
l
t
a
r

x-ai/grok-3-mini-beta

x-ai

Failed

c
r
a
n
e
q
u
a
i
l
w
a
l
k
s
p
o
l
k
a
h
a
l
a
l
d
a
l
l
y

mistralai/mistral-medium-3

mistral

Failed

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

anthropic/claude-3.7-sonnet

anthropic

Failed

c
r
a
n
e
b
l
o
a
t
f
l
o
y
t
p
l
o
u
t
g
l
o
u
t
a
l
o
f
t

qwen/qwen3-235b-a22b

qwen

2/6

c
r
a
n
e
a
d
o
p
t
a
b
b
o
t
a
f
o
o
t
a
l
l
o
t

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