AI vs Puzzles

Back to Home
Puzzle#1697

Wordle

Released: 2026-02-10

8
Models Tested
5
Solved
5.00
Avg. Attempts
8
LLM Providers
Correct
Present
Absent

google/gemini-2.0-flash-001

google

2/6

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

openai/gpt-4o

openai

2/6

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

anthropic/claude-3.7-sonnet

anthropic

Failed

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

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

meta

Failed

h
o
u
s
e
b
l
a
z
e
s
p
a
c
e
s
c
a
r
e
s
c
o
p
e
s
l
i
c
e

deepseek/deepseek-chat-v3-0324

deepseek

Failed

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

x-ai/grok-3-mini-beta

x-ai

2/6

c
r
a
n
e
s
c
e
n
e

qwen/qwen3-235b-a22b

qwen

2/6

c
r
a
n
e
s
c
o
n
e
s
c
e
n
e

mistralai/mistral-medium-3

mistral

2/6

c
r
a
n
e
u
n
m
e
t
n
i
c
e
r
d
a
n
c
e
c
o
n
e
s
s
c
e
n
e

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