AI vs Puzzles

Back to Home
Puzzle#1695

Wordle

Released: 2026-02-08

8
Models Tested
4
Solved
5.63
Avg. Attempts
8
LLM Providers
Correct
Present
Absent

mistralai/mistral-medium-3

mistral

2/6

c
r
a
n
e
e
v
a
d
e
e
d
g
e
d
e
e
r
i
e
e
j
e
c
t
e
m
b
e
d

google/gemini-2.0-flash-001

google

Failed

r
a
i
s
e
c
l
o
n
e
f
u
d
g
e
t
h
e
m
e
b
l
e
e
p
e
m
b
e
r

deepseek/deepseek-chat-v3-0324

deepseek

2/6

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

openai/gpt-4o

openai

2/6

c
r
a
n
e
s
w
i
p
e
b
r
o
k
e
b
l
e
e
d
d
e
b
e
t
e
m
b
e
d

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

meta

Failed

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

x-ai/grok-3-mini-beta

x-ai

Failed

c
r
a
n
e
p
i
x
e
l
h
o
s
e
d
f
e
t
e
d
d
u
k
e
d
d
e
w
e
d

anthropic/claude-3.7-sonnet

anthropic

Failed

s
l
a
t
e
p
r
o
n
e
g
u
i
d
e
c
h
y
d
e
w
b
k
e
d
v
e
x
e
d

qwen/qwen3-235b-a22b

qwen

2/6

c
r
a
n
e
f
l
e
e
t
d
e
w
e
d
e
g
g
e
d
e
m
b
e
d

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