SimmonsBench — compare
COMPARE
Two models, head to head.
Pick two — their builds side by side, and where they land on every board.
VS
The builds
Same prompt, both models. 60 tests — click any build to open it.
FILTER
Opus 4.8
◉GPT-5.5
DASHBOARD
Build a financial spend dashboardSLIDES
Build a sales QBR deckDASHBOARD
Turn raw sales data into a dashboardDATA-VIZ
Build an interactive chartWRITING
Write a cold outreach emailDASHBOARD · DATA
Build a make-vs-buy modelDASHBOARD · DATA
Model 3 pricing strategiesREPORT · DATA
Synthesize survey + analytics into a reportMAP · DATA
Build an interactive earthquake mapUI · IMAGE
Design a mobile app UICANVAS
Build a physics sandboxCANVAS
Build a solar systemWEBGL · AGENTIC
Build a 3D art museumWRITING
Write a viral tweet threadWRITING
Write a viral LinkedIn postVIDEO · AGENTIC
Make a launch video for yourself (HyperFrames)IMAGE · STORYBOARD
Pitch the next A24 film (+ storyboard)WRITING · LONG-FORM
Finish a year-old essaybuild.exe×
no build…
WRITING · EXPLAINER
Explain how an AI model worksbuild.exe×
no build…
WRITING · SOCIAL
Write a LinkedIn post (to a system prompt)build.exe×
no build…
WRITING · VOICE-MATCH
Write a YouTube intro in a creator's voicebuild.exe×
no build…
GAME · GRAPHICS
Build a great-looking racing gamebuild.exe×
no build…
WEBSITE · DESIGN TASTE
Design a freelance designer's portfolio sitebuild.exe×
no build…
3D · WEBGL
Build a 3D nebula galaxy scenebuild.exe×
no build…
3D · PROCEDURAL
Build an endless low-poly terrain flyoverbuild.exe×
no build…
SIM · FLUID
Animate liquid sloshing in a spherebuild.exe×
no build…
build.exe×
no build…
CANVAS · MATH
Build an interactive fractal explorerbuild.exe×
no build…
build.exe×
no build…
3D · LOGIC
Build a Rubik's Cube solverbuild.exe×
no build…
build.exe×
no build…
build.exe×
no build…
build.exe×
no build…
DATA-VIZ · 3D
Build an interactive Milky Way learning mapbuild.exe×
no build…
build.exe×
no build…
GAME · WEBGL
Build a Counter-Strike 2 browser clonebuild.exe×
no build…
build.exe×
no build…
DATA · GRAPH
Build a cinema knowledge graphbuild.exe×
no build…
build.exe×
no build…
build.exe×
no build…
build.exe×
no build…
build.exe×
no build…
build.exe×
no build…
WRITING · VOICE
Write a YouTube intro in the creator's voicebuild.exe×
no build…
build.exe×
no build…
build.exe×
no build…
build.exe×
no build…
build.exe×
no build…
build.exe×
no build…
3D · INTERACTIVE
Build an interactive 3D command-center gadgetbuild.exe×
no build…
build.exe×
no build…
LIVE DATA · FORECAST
Build a weather app with its own forecast modelbuild.exe×
no build…
build.exe×
no build…
ALGORITHM · AUDIO
Build an MP3-to-MIDI converterbuild.exe×
no build…
build.exe×
no build…
build.exe×
no build…
build.exe×
no build…
WRITING · HALLUCINATION TEST
Write an Isaacson-style biography introbuild.exe×
no build…
build.exe×
no build…
WRITING · AI VIDEO
Write + shot-list a 30-second Tarantino trailerbuild.exe×
no build…
build.exe×
no build…
MOTION · VIDEO
Author a 20-second motion-graphics explainerbuild.exe×
no build…
build.exe×
no build…
3D · PRODUCT PAGE
Build a luxury watch product page (3D)build.exe×
no build…
build.exe×
no build…
KNOWLEDGE WORK · COPY
Rewrite a real workshop landing pagebuild.exe×
no build…
build.exe×
no build…
● The Field
SimmonsBench
Leaderboards
Composite · all 7 boards · scored out of 100
SimmonsBench
1
Opus 4.8
12.16 /100
2
Fable 5
11.51 /100
3
GPT-5.5
10.70 /100
4
Gemini 3.5
10.66 /100
5
Grok 4.3
9.02 /100
6
GLM 5.2
7.83 /100
7
DeepSeek V4
7.72 /100
8
Kimi K2.7
7.59 /100
9
Qwen 3.7
7.42 /100
10
Mistral Large
7.07 /100
11
MiniMax M3
6.31 /100
12
GPT-OSS 120b
5.90 /100
13
Sonnet 5
0.00 /100
14
Sonnet 4.6
0.00 /100
See the full ranking + how it's scored
→