multimodal benchmark
LifeSciBench
LifeSciBench is an expert-authored, expert-reviewed benchmark of 750 open-ended life-science research tasks spanning seven workflows and seven biological domains. Responses are graded against 19,020 physician- and scientist-written rubric criteria rather than multiple-choice answers, and most tasks require interpreting attached artifacts such as figures, PDFs, and sequence files.
Updated Aug 11, 2026
Models3
Model coverage3
MetricScore
EvidenceB
LifeSciBench Ranking
Higher score ranks better on this benchmark.
| 1 | OP | 59.9% | 100.0% | 3 | C | |
| 2 | OP | 56.0% | 50.0% | 3 | C | |
| 3 | OP | 51.2% | 0.0% | 3 | C |
LifeSciBench Score Distribution
A closer view of the leading scores on this benchmark.
LifeSciBench Highlights
The leading models and scores on this benchmark.
What is LifeSciBench?
What LifeSciBench measures and how its scores work.
LifeSciBench is an expert-authored, expert-reviewed benchmark of 750 open-ended life-science research tasks spanning seven workflows and seven biological domains. Responses are graded against 19,020 physician- and scientist-written rubric criteria rather than multiple-choice answers, and most tasks require interpreting attached artifacts such as figures, PDFs, and sequence files.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- LifeSciBench
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- lifescibench|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about LifeSciBench.
Which model scores highest on LifeSciBench?
GPT-5.6 Sol is currently ranked first with 59.9%.
What does LifeSciBench measure?
LifeSciBench is an expert-authored, expert-reviewed benchmark of 750 open-ended life-science research tasks spanning seven workflows and seven biological domains. Responses are graded against 19,020 physician- and scientist-written rubric criteria rather than multiple-choice answers, and most tasks require interpreting attached artifacts such as figures, PDFs, and sequence files.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
3 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.