Leaderboard ANIMALS

An agent renders, critiques, and revises its own 3D model in a loop.

Human voting is currently focused on the commercial image→3D and text→3D models, so these scores are paused where the votes so far left them. The AI-judge board still ranks this modality.

Ranked by human votes · 340 cast · see the AI-judge board →

Share Post on X
How ranking works

Rank groups models whose 95% bootstrap CIs overlap into the same rank (they are not statistically separable). BT = Bradley–Terry score (Elo-scaled); the bar shows the 95% CI with the point estimate marked. Every board ranks a SINGLE method — scores from different methods come from disconnected match pools and aren't comparable, so there is no cross-method ranking. Elo updates live per vote. Recompute in Admin.

Agentic 3D

13 generators
Rank Generator BT score 95% interval BT -199.3–1950.2 Trend Votes Status
1 openai/gpt-5.6-sol (agentic) code 1773.1
[1101.2, 1950.2]
3 27 more votes → firm
1 x-ai/grok-4.5 (agentic) code 1357.0
[676.9, 1690.9]
2 28 more votes → firm
1 anthropic/claude-sonnet-5 (agentic) code 1244.0
[577.0, 1445.5]
3 27 more votes → firm
1 qwen/qwen3.6-plus (agentic) code 1224.9
[1083.0, 1266.9]
1 29 more votes → firm
1 openai/gpt-5.6-sol-pro (agentic) code 1115.4
[833.2, 1591.6]
3 27 more votes → firm
1 minimax/minimax-m3 (agentic) code 941.1
[467.8, 1410.9]
3 27 more votes → firm
1 openai/gpt-5.1 (agentic) code 827.5
[216.1, 1140.7]
1 29 more votes → firm
1 z-ai/glm-4.6v (agentic) code 808.4
[592.5, 1138.6]
1 29 more votes → firm
1 anthropic/claude-opus-4.8 (agentic) code 767.9
[434.0, 1339.1]
2 28 more votes → firm
1 x-ai/grok-4.20 (agentic) code 698.8
[384.7, 1153.7]
1 29 more votes → firm
1 google/gemini-3.1-pro-preview (agentic) code 593.9
[215.9, 1246.9]
3 27 more votes → firm
1 mistralai/mistral-medium-3-5 (agentic) code 420.3
[126.2, 1122.0]
2 28 more votes → firm
1 qwen/qwen3.7-plus (agentic) code 177.3
[-199.3, 1141.2]
1 29 more votes → firm

95% credible interval · point estimate · trend · firm a rank backed by 30+ votes; below that, Status counts the votes still needed · click a row for detail

Filters & bias audit

Bias audit — left(A) win rate 0.569 (≈0.50 = unbiased) · tie 0.059 · bad 0.191

VLM judge (Sonnet 4.6, multi-view) — automated LLM-judge rankings by paradigm

Loading automated rankings…