Atlas

Benchmarks

← All benchmarks

BuseyBench SVG

Multimodal · 2026-07-02

BuseyBench SVG measures how well a model creates a standalone 1024×1024 SVG portrait of Gary Busey from one fixed no-web prompt. Three vision judges each score the rendered output three times; BuseyBench takes each judge's median and then averages the judges. Its raw 0–10 score combines Busey likeness (50%), face coherence (20%), visual aesthetics (15%), and prompt adherence (15%); Atlas stores this raw absolute score rather than the leaderboard's pairwise-adjusted ranking score.

Top models (higher is better)

ModelScore
GPT-5.6 Terra Pro6.5
GPT-5.6 Sol Pro6.3
Claude Opus 56.3
Gemini 3.1 Pro Preview6.3
Kimi K36.2
GPT-5.6 Sol6.1
GPT-5.6 Terra6.1
GPT-5.5 Pro6.1
GPT-5.56.0
Gemini 3.5 Flash5.9
Claude Fable 55.9
Fugu Ultra v1.05.8
GPT-5.6 Luna Pro5.7
Opus 4.65.5
GPT-5.2 Pro5.5
Nex-N2-Pro5.5
Opus 4.85.4
Kimi K2.65.2
Opus 4.55.2
GPT-5.2-Codex5.2
Qwen3.7-Plus5.1
Opus 4.75.1
GLM 5V Turbo5.1
GPT-5.25.1
Gemini 3.6 Flash5.0
GLM-5.25.0
Qwen3.5 Plus (2026-02-15)4.9
Grok 4.54.9
GPT-5.1-Codex-Max4.9
GPT-5.3-Codex4.9
Sonnet 4.64.8
GPT-5.3 Instant4.8
Sonar Pro4.7
Muse Spark 1.14.7
Qwen3.7-Max4.6
o34.6
GPT-5.44.6
Sonnet 54.5
GPT-54.5
Qwen3.6 27B4.5
Loading Atlas data…