AI vs Puzzles

Back to Home
Puzzle#1693

Wordle

Released: 2026-02-06

8
Models Tested
1
Solved
5.75
Avg. Attempts
8
LLM Providers
Correct
Present
Absent

mistralai/mistral-medium-3

mistral

Failed

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

google/gemini-2.0-flash-001

google

Failed

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

openai/gpt-4o

openai

Failed

c
r
a
t
e
l
e
a
s
e
b
l
a
m
e
l
e
a
f
y
e
q
u
a
l
m
e
d
a
l

anthropic/claude-3.7-sonnet

anthropic

Failed

a
r
i
s
e
c
a
m
e
l
b
a
g
e
l
p
a
g
e
l
h
a
g
e
l
k
a
g
e
l

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

meta

Failed

h
o
u
s
e
f
l
a
m
e
b
a
d
g
e
l
a
g
e
r
e
l
a
g
e
a
n
g
l
e

deepseek/deepseek-chat-v3-0324

deepseek

Failed

c
r
a
n
e
p
l
a
t
e
b
l
a
m
e
f
l
a
m
e
g
l
a
z
e
g
a
b
l
e

qwen/qwen3-235b-a22b

qwen

2/6

c
r
a
n
e
s
w
e
a
t
b
a
g
e
l
g
a
v
e
l

x-ai/grok-3-mini-beta

x-ai

Failed

c
r
a
n
e
s
t
e
a
k
f
a
d
e
d
h
a
z
e
l
l
a
p
e
l
b
a
g
e
l

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