BREAKING
OpenAI Unveils LifeSciBench
0
contributors
0
tasks
0
workflows
How tasks are scored
1
Correctness
↓
2
Scientific validity
↓
3
Usefulness
↓
4
Justification
↓
5
Caveats
0
reviewers
0
artifacts
0
rubric criteria
GPT-Rosalind vs GPT-5.5
MedChemBench R
27.5
MedChemBench 5.5
25.1
LabWork R
63.2
LabWork 5.5
55.8
A new bar for AI in science
AI NEWS BLITZ
OpenAI has released LifeSciBench, a new benchmark for AI in life science research.