Healthcare, biotechnology & medical AI Research
Med42-v2 studies medical question answering in two Llama 3 configurations
M42’s Med42-v2 paper documents 8B and 70B Llama 3 variants evaluated on medical-language benchmarks. [1][2]
What the researchers found
The authors submitted the Med42-v2 preprint on August 12, 2024. M42’s model card identifies the separately documented 70B variant. [1][2]
Why it matters
Reporting both sizes gives researchers a way to compare model behavior on the same fixed language tasks. [1][2]
Who can use the findings
The research is relevant to people evaluating language-model performance on medical question-answering datasets. [1][2]
Key dates
- M42 authors submit the Med42-v2 preprint [1]
Practical uses
The work reports empirical benchmark comparisons with other language models under the authors’ test setup. [1][2]
Limits of the evidence
The benchmark report concerns research test performance. It does not establish safe clinical deployment or a regulatory clearance. [1][2]
Sources and verification
Catalog evidence checked How we verify records
- Primary source M42 authors’ original Med42-v2 research preprint
- Primary source m42-health/Llama3-Med42-70B
- Other source Med42-v2: A Suite of Clinical LLMs
Related profiles
Healthcare, biotechnology & medical AI
Sleep Analysis Module (SAM)
Neurotronics (Nihon Kohden)
The module estimates sleep time from a signal captured during its intended sleep-testing workflow.
Healthcare, biotechnology & medical AI
Heartvue.Proton
Heartvue.ai
It is software for analyzing cardiac MRI images within its documented intended-use scope.
Healthcare, biotechnology & medical AI
ThinkSono Guidance
ThinkSono
Provides guidance during acquisition of lower-extremity compression ultrasound images for review by healthcare professionals.