Phylo
BiomniBench-DataAnalysis (Biomni-DA-v0)
Preliminary 15-task preview of Phylo's trace-based evaluation for long-horizon biological data analysis. It scores analytical process—including data handling, method selection, statistical rigor, source reliability, and reasoning—rather than only final-answer accuracy. Results are preliminary and the benchmark remains a work in progress.
biology-researchbiomedical-data-analysistrace-evaluationtabularcodetext
Status
curated
Saturation
unsaturated
Reproduction
not yet
Published results
6 models
Editor's note
Preliminary 15-task preview of Phylo's trace-based evaluation for long-horizon biological data analysis. It scores analytical process—including data handling, method selection, statistical rigor, source reliability, and reasoning—rather than only final-answer accuracy. Results are preliminary and the benchmark remains a work in progress.
Published results
| Model | Score | Variant | Provenance |
|---|---|---|---|
| Senior pharmaceutical scientists | 68.5% trace score | human comparison group | external |
| Biomni Lab (2026-02-03) | 65.0% trace score | Biomni-DA-v0 · 15 tasks | external |
| Edison Analysis | 52.7% trace score | Biomni-DA-v0 · 15 tasks | external |
| OpenAI Agents SDK (GPT-5.2) | 51.4% trace score | Biomni-DA-v0 · 15 tasks | external |
| Junior scientists | 48.5% trace score | human comparison group | external |
| Claude Code (Opus 4.6) | 47.8% trace score | Biomni-DA-v0 · 15 tasks | external |
Sources
- external Phylo — BiomniBench preliminary results