Atlas

Benchmarks

← All benchmarks

GDPx: Biology

Science · 2026-11-01

GDPx's Biology domain: 13 tasks from LatchBio on wet-lab protocol design, sequence analysis and computational biology pipelines. Scores are pass@1 over three attempts per task. A component of the GDPx score, which weights its 11 domains equally.

Top models (higher is better)

ModelScore
Gemini 4 Carbon25.6
Claude Fable 5.120.5
Claude Opus 5.520.5
Muse Wonder17.9
GPT-6 Luna15.4
Gemini 4 Argon12.8
GPT-6.1 Terra12.8
Muse Spark Open12.8
Claude Haiku 5.57.7
Grok 4.77.7
MiniMax M3.1 Flash7.7
Claude Sonnet 5.55.1
GPT-6 Astra5.1
Kimi K35.1
Muse Wonder Full5.1
GPT-6.1 Sol2.6
GLM-5.30.0
Loading Atlas data…