AI vs Puzzles

Back to Home
Puzzle#1696

Wordle

Released: 2026-02-09

8
Models Tested
3
Solved
5.50
Avg. Attempts
8
LLM Providers
Correct
Present
Absent

mistralai/mistral-medium-3

mistral

2/6

c
r
a
n
e
c
i
v
i
c
c
l
o
m
p
c
l
e
r
k
c
l
e
a
t
c
e
l
l
o

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

meta

Failed

h
o
u
s
e
p
h
o
t
o
c
l
o
a
k
c
l
o
s
e
c
o
l
o
r
c
l
o
n
e

openai/gpt-4o

openai

Failed

p
l
a
n
e
l
i
s
l
e
m
o
d
e
l
o
l
i
v
e
e
l
b
o
w
g
l
o
v
e

anthropic/claude-3.7-sonnet

anthropic

Failed

r
a
i
s
e
c
l
o
v
e
c
o
d
l
e
c
h
o
l
e
c
o
b
l
e
c
o
p
l
e

google/gemini-2.0-flash-001

google

Failed

r
a
i
s
e
c
l
o
n
e
c
i
v
i
l
c
o
u
l
d
c
o
l
o
r
c
h
a
l
k

deepseek/deepseek-chat-v3-0324

deepseek

Failed

c
r
a
n
e
c
o
v
e
y
c
o
m
e
t
c
h
o
k
e
c
o
v
e
y
c
o
p
e
r

x-ai/grok-3-mini-beta

x-ai

2/6

c
r
a
n
e
c
h
e
s
t
c
l
u
e
d
c
e
l
l
o

qwen/qwen3-235b-a22b

qwen

2/6

r
a
i
s
e
b
e
l
o
w
h
e
l
l
o
c
e
l
l
o

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