Scoreboard · 25 points

The leaderboard

Eight runs, sorted by total score. Each total is the sum of five editorial scores, each out of 5.

Tested · 2026-10-27 Consumer client, zero-shot 8 runs
The leaderboard
Rank Bib Model SemanticsAnatomyBicycleMotionPolish Total
1 03 Doubao 2.1 ProByteDance 5.04.05.04.55.0 23.5/25
2 02 Kimi K3.1 ProMoonshot AI 5.03.54.04.05.0 21.5/25
3 04 GPT-6.1 SolOpenAI 4.53.04.04.54.5 20.5/25
4 07 GPT-6.0 LunaOpenAI 4.53.04.02.54.5 18.5/25
5 01 DeepSeek FlashDeepSeek 4.53.55.01.53.5 18.0/25
6 06 GPT-5.6 SolOpenAI 4.53.53.01.55.0 17.5/25
7 08 HY3Not stated 4.52.53.51.54.0 16.0/25
8 05 Agnes 3.0 FlashAgnes AI 0.51.03.52.02.5 9.5/25

One zero-shot run per model, consumer clients, 27 Oct 2026. These are editorial scores, not an automated or general-intelligence ranking. Client-side model versions can change without notice.