⏱ Capsule Public Vibe Design

Lighthouse at Dusk: Free-Style SVG

Ezarwebmaster· Sep 29, 2026

Locked Reference Prompt

IMMUTABLE

Scientific timeline lock active

Draw a lighthouse at dusk on a rocky cliff above the sea, as a single SVG. Requirements: - Output ONLY the raw SVG code, no explanation, no markdown fences. - viewBox="0 0 800 800", no external resources, no <image>, no JavaScript. - Make it as beautiful and polished as you can. Every artistic choice (style, palette, composition, lighting, details) is yours. - Sign your work: add your own model name (the exact name of the model generating this SVG) as visible text in a <text> element, in a discreet corner. If you are unsure of your exact name, write your best guess.

Add a Benchmark Run

Sign in to run this prompt against hundreds of models with your own OpenRouter key and archive the results.

Timeline (11 runs)

Run Activity

11 runs in the last 6 months

Apr
May
Jun
Jul
Aug
Sep
LessMore

Cost vs Speed

size = output tokens · top-left is best

100$0.010$0.10fast · cheapfast · priceyslow · cheapslow · priceyCost per request ($, log)Speed (tokens/sec, log)
size = tokens4.2k17k42k
value frontier — no model is faster & cheaper

Capsule Stats

Runs
11
Total cost
$1.1716
Tokens
165k
Avg latency
2m 17s
Models
11
Web Searches
0
Top provider Anthropic×3
57k reasoning last today
Tip: Select 2 or more runs via their "Compare" buttons — or filter by company below and compare them all at once — then open the Compare Studio: verdicts, benchmark bars, charts, side-by-side reading and response diff.
⏱

Benchmark Run — Sep 29, 2026 EzarwebmasterEzarwebmaster Latest

MiniMax
MiniMaxminimax-m3
🌐 Web Search (Not Used)
minimax/minimax-m3Sep 29, 2026, 10:58:11 PM
Possible issue
length
Latency
2m 33s
client → response
Input Tokens
904
prompt tokens
Output
4,096
generated
Total Tokens
5,000
in + out
Billed Cost
$0.0040
OR Credits
Reasoning
3,142
thinking tokens
Model Output

(No text response)

This reasoning model spent its entire output budget on thinking (3,142 tokens) and hit the limit before writing any answer. Use the retry button by the LENGTH badge to re-run with a higher cap.

⏱

Benchmark Run — Sep 29, 2026 EzarwebmasterEzarwebmaster

NVIDIA
NVIDIAnemotron-3-ultra-550b-a55b
🌐 Web Search (Not Used)
nvidia/nemotron-3-ultra-550b-a55bSep 29, 2026, 10:54:34 PM
Extended
stop
Latency
21s
client → response
Input Tokens
780
prompt tokens
Output
5,241
generated
Total Tokens
6,021
in + out
Billed Cost
$0.0119
OR Credits
Reasoning
179
thinking tokens
Model Output
nemotron-3-ultra-550b-a55b — SVG
nemotron-3-ultra-550b-a55b — SVG
⏱

Benchmark Run — Sep 29, 2026 EzarwebmasterEzarwebmaster

Anthropic
Anthropicclaude-fable-5.1
🌐 Web Search (Not Used)
anthropic/claude-fable-5.1Sep 29, 2026, 10:14:46 PM
Extended
stop
Latency
1m 19s
client → response
Input Tokens
2,941
prompt tokens
Output
6,765
generated
Total Tokens
9,706
in + out
Billed Cost
$0.3677
OR Credits
Reasoning
—
thinking tokens
Model Output
claude-fable-5.1 — SVG
claude-fable-5.1 — SVG
⏱

Benchmark Run — Sep 29, 2026 EzarwebmasterEzarwebmaster

OpenAI
OpenAIgpt-6-astra
🌐 Web Search (Not Used)
openai/gpt-6-astraSep 29, 2026, 10:14:38 PM
Extended
stop
Latency
1m 55s
client → response
Input Tokens
4,579
prompt tokens
Output
6,133
generated
Total Tokens
10,712
in + out
Billed Cost
$0.3637
OR Credits
Reasoning
136
thinking tokens
Model Output
gpt-6-astra — SVG
gpt-6-astra — SVG
⏱

Benchmark Run — Sep 29, 2026 EzarwebmasterEzarwebmaster

Qwen (Alibaba)
Qwen (Alibaba)qwen3.8-omni-flash
🌐 Web Search (Not Used)
qwen/qwen3.8-omni-flashSep 29, 2026, 10:13:03 PM
Extended
stop
Latency
6m 40s
client → response
Input Tokens
814
prompt tokens
Output
42,249
generated
Total Tokens
43,063
in + out
Billed Cost
$0.0200
OR Credits
Reasoning
27,585
thinking tokens
Model Output
qwen3.8-omni-flash — SVG
qwen3.8-omni-flash — SVG
⏱

Benchmark Run — Sep 29, 2026 EzarwebmasterEzarwebmaster

Qwen (Alibaba)
Qwen (Alibaba)qwen3.8-flash
🌐 Web Search (Not Used)
qwen/qwen3.8-flashSep 29, 2026, 10:12:41 PM
Extended
stop
Latency
5m 31s
client → response
Input Tokens
823
prompt tokens
Output
24,435
generated
Total Tokens
25,258
in + out
Billed Cost
$0.0116
OR Credits
Reasoning
10,043
thinking tokens
Model Output
qwen3.8-flash — SVG
qwen3.8-flash — SVG
⏱

Benchmark Run — Sep 29, 2026 EzarwebmasterEzarwebmaster

Anthropic
Anthropicclaude-sonnet-5.5
🌐 Web Search (Not Used)
anthropic/claude-sonnet-5.5Sep 29, 2026, 10:06:53 PM
Extended
stop
Latency
48s
client → response
Input Tokens
3,024
prompt tokens
Output
7,553
generated
Total Tokens
10,577
in + out
Billed Cost
$0.0816
OR Credits
Reasoning
—
thinking tokens
Model Output
claude-sonnet-5.5 — SVG
claude-sonnet-5.5 — SVG
⏱

Benchmark Run — Sep 29, 2026 EzarwebmasterEzarwebmaster

Anthropic
Anthropicclaude-opus-5.5
🌐 Web Search (Not Used)
anthropic/claude-opus-5.5Sep 29, 2026, 10:06:46 PM
Extended
stop
Latency
1m 3s
client → response
Input Tokens
2,941
prompt tokens
Output
6,703
generated
Total Tokens
9,644
in + out
Billed Cost
$0.1458
OR Credits
Reasoning
1,177
thinking tokens
Model Output
claude-opus-5.5 — SVG
claude-opus-5.5 — SVG
⏱

Benchmark Run — Sep 29, 2026 EzarwebmasterEzarwebmaster

OpenAI
OpenAIgpt-6.1-sol
🌐 Web Search (Not Used)
openai/gpt-6.1-solSep 29, 2026, 10:06:37 PM
Extended
stop
Latency
1m 56s
client → response
Input Tokens
4,579
prompt tokens
Output
6,997
generated
Total Tokens
11,576
in + out
Billed Cost
$0.0814
OR Credits
Reasoning
209
thinking tokens
Model Output
gpt-6.1-sol — SVG
gpt-6.1-sol — SVG
⏱

Benchmark Run — Sep 29, 2026 EzarwebmasterEzarwebmaster

Google DeepMind
Google DeepMindgemini-3.8-flash
🌐 Web Search (Not Used)
google/gemini-3.8-flashSep 29, 2026, 10:06:26 PM
Extended
stop
Latency
2m 24s
client → response
Input Tokens
169
prompt tokens
Output
17,426
generated
Total Tokens
17,595
in + out
Billed Cost
$0.0655
OR Credits
Reasoning
4,480
thinking tokens
Model Output
gemini-3.8-flash — SVG
gemini-3.8-flash — SVG
⏱

Benchmark Run — Sep 29, 2026 EzarwebmasterEzarwebmaster

DeepSeek
DeepSeekdeepseek-v4.1-flash
🌐 Web Search (Not Used)
deepseek/deepseek-v4.1-flashSep 29, 2026, 10:05:25 PM
Extended
stop
Latency
36s
client → response
Input Tokens
797
prompt tokens
Output
15,148
generated
Total Tokens
15,945
in + out
Billed Cost
$0.0184
OR Credits
Reasoning
9,857
thinking tokens
Model Output
deepseek-v4.1-flash — SVG
deepseek-v4.1-flash — SVG