rankings

Coding arena

Coding Arena

One coding prompt, given to every model word for word. Each card runs that model's answer, live and sandboxed. No edits, no fixes. What it shipped is what you click.

The prompt every model gottask vglass-fill-1.1.0show

Build a 2D game in a single self-contained HTML file (HTML + CSS + JS, no external libraries, no network). The game must be 1000x1000 pixels. On screen is an empty glass. Each click pours water in; the glass fills up over repeated clicks, with a visible animated water level. Include a way to empty/reset it. Keep it to one file that runs by just opening it in a browser.

Sort

anthropic/claude-opus-4-5

anthropic/claude-opus-4-5
requestyRrequesty
loading…
tokens4,169207→3,962
cost$0.0042$1/M
time44.2s
code384lines
score 0

anthropic/claude-opus-4-6

anthropic/claude-opus-4-6
requestyRrequesty
loading…
tokens7,527207→7,320
cost$0.0075$1/M
time86.7s
code640lines
score 0

openai/gpt-5.5

openai/gpt-5.5
requestyRrequesty
loading…
tokens4,660169→4,491
cost$0.0047$1/M
time63.6s
code528lines
score 0

anthropic/claude-sonnet-4-5

anthropic/claude-sonnet-4-5
requestyRrequesty
loading…
tokens3,392207→3,185
cost$0.0034$1/M
time40.3s
code345lines
score 0

anthropic/claude-opus-4-8

anthropic/claude-opus-4-8
requestyRrequesty
loading…
tokens6,013241→5,772
cost$0.0060$1/M
time57.6s
code445lines
score 0

anthropic/claude-opus-4-7

anthropic/claude-opus-4-7
requestyRrequesty
loading…
tokens7,524246→7,278
cost$0.0075$1/M
time75.9s
code529lines
score 0

anthropic/claude-sonnet-5

anthropic/claude-sonnet-5
requestyRrequesty
loading…
tokens5,584241→5,343
cost$0.0056$1/M
time42.0s
code358lines
score 0

anthropic/claude-fable-5

anthropic/claude-fable-5
requestyRrequesty
loading…
tokens8,414241→8,173
cost$0.0084$1/M
time85.8s
code427lines
score 0

mistral/mistral-large-latest

mistral/mistral-large-latest
requestyRrequesty
loading…
tokens977178→799
cost$0.0010$1/M
time9.5s
code120lines
score 0

anthropic/claude-sonnet-4-6

anthropic/claude-sonnet-4-6
requestyRrequesty
loading…
tokens7,615207→7,408
cost$0.0076$1/M
time92.9s
code645lines
score 0

google/gemini-3.1-pro-preview

google/gemini-3.1-pro-preview
requestyRrequesty
loading…
tokens5,838167→5,671
cost$0.0058$1/M
time41.4s
code210lines
score 0

openai/gpt-5.2-chat

openai/gpt-5.2-chat
requestyRrequesty
loading…
tokens986169→817
cost$0.0010$1/M
time7.1s
code138lines
score 0

openai/gpt-5.4

openai/gpt-5.4
requestyRrequesty
loading…
tokens4,592169→4,423
cost$0.0046$1/M
time25.5s
code562lines
score 0

deepinfra/Qwen/Qwen3-32B

deepinfra/Qwen/Qwen3-32B
requestyRrequesty
loading…
tokens7,125176→6,949
cost$0.0071$1/M
time77.6s
code786lines
score 0

openai/gpt-5.4-mini

openai/gpt-5.4-mini
requestyRrequesty
loading…
tokens4,830169→4,661
cost$0.0048$1/M
time22.8s
code519lines
score 0

openai/gpt-5-mini

openai/gpt-5-mini
requestyRrequesty
loading…
tokens6,140169→5,971
cost$0.0061$1/M
time45.1s
code593lines
score 0

xai/grok-4-fast

xai/grok-4-fast
requestyRrequesty
loading…
tokens2,204292→1,912
cost$0.0022$1/M
time14.1s
code117lines
score 0

openai/gpt-5.2

openai/gpt-5.2
requestyRrequesty
loading…
tokens7,143169→6,974
cost$0.0071$1/M
time70.8s
code803lines
score 0

alibaba/qwen3.6-plus

alibaba/qwen3.6-plus
requestyRrequesty
loading…
tokens6,575178→6,397
cost$0.0066$1/M
time357.6s
code615lines
score 0

minimaxi/MiniMax-M2.7

minimaxi/MiniMax-M2.7
requestyRrequesty
loading…
tokens7,063176→6,887
cost$0.0071$1/M
time92.8s
code257lines
score 0

openai/gpt-5.6-luna

openai/gpt-5.6-luna
requestyRrequesty
loading…
tokens2,906169→2,737
cost$0.0029$1/M
time14.4s
code305lines
score 0

openai/gpt-5.6-terra

openai/gpt-5.6-terra
requestyRrequesty
loading…
tokens3,583169→3,414
cost$0.0036$1/M
time21.9s
code347lines
score 0

openai/gpt-5.6-sol

openai/gpt-5.6-sol
requestyRrequesty
loading…
tokens5,242169→5,073
cost$0.0052$1/M
time61.6s
code522lines
score 0

Each game runs in a sandboxed iframe (scripts only, no network, no access to this page). The numbers were measured when the game was recorded: tokens (input→output), cost (usage × the fixed price table), time (how long the model took), and code (raw line count).