A same-condition pelican archive for comparing versions and reasoning efforts — entertainment and reference only

Independent hobby project. Not affiliated with or endorsed by Simon Willison, pelicanbenchmark.com, or Pelican Zoo.

In July 2026 Simon Willison wrote that the correlation between pelican scores and real model quality “has been mostly severed now”

Why this site is an archive for family comparison →

Recently added

Demo Pelican · 2026-09 · default

Demo Pelican · 2026-09 · low

Demo Pelican · 2026-09 · high

Example Model · 2026-08 · default

Example Model · 2026-08 · low

Example Model · 2026-08 · high

Mock Wing · 2026-07 · default

Sample Bird · 2026-06 · default

Reference board

The AI reference score comes from AI judges, has known bias, and is not a measure of overall model ability.

Selected by AI reference score

AI reference score Visitor votes (Phase 2) · Coming soon Show all efforts Latest version only ExampleAI DemoLabs Placeholder Inc. From To

		Thumb	Model	Vendor	Version	Effort	Release date	Reference score	Valid SVG rate	Median duration	Samples	Elo
			Demo Pelican	DemoLabs	2026-09	default	2026-09-12	8.4 ± 0.3	100%	19.2 s	5/5	Coming soon
	Monitoring · M3 · 2026-10-08		Example Model	ExampleAI	2026-08	default	2026-08-20	7.7 ± 0.3	100%	9.7 s	5/5	Coming soon
			Example Model	ExampleAI	2026-03	default	2026-03-15	6.7 ± 0.3	100%	6.5 s	5/5	Coming soon
			Sample Bird	Placeholder Inc.	2026-06	default	2026-06-01	6.5 ± 0.3	100%	8.0 s	5/5	Coming soon
			Demo Pelican	DemoLabs	2026-04	default	2026-04-08	5.7 ± 0.2	100%	16.0 s	5/5	Coming soon
			Example Model	ExampleAI	2025-11	default	2025-11-01	4.8 ± 0.2	100%	5.0 s	5/5	Coming soon
			Demo Pelican	DemoLabs	2025-12	default	2025-12-10	4.6 ± 0.2	100%	14.5 s	5/5	Coming soon
	No longer in Cursor		Mock Wing	ExampleAI	2026-07	default	2026-07-18	3.6 ± 0.3	100%	9.6 s	5/5	Coming soon

Insufficient samples (unranked by default)

Insufficient samples

Fictional Beak

DemoLabs

2026-05

default

2026-05-22

7.5 ± 0.0

40%

11.1 s

2/5

Coming soon