MedConceptsQA
MedConceptsQA is a dedicated open source benchmark for medical concepts question answering.
How to run
ltb · medconceptsqa
$ npm i -g lettokenburn-cli$ ltb bench pull medconceptsqa --n 50$ ltb bench run ./medconceptsqa -a claude-code -m anthropic/claude-opus-4-8$ ltb bench push
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
Scores
3 sourced entries| Harness | Source | Confidence | Link | |||
|---|---|---|---|---|---|---|
| GPT-4 0125 Preview | 61.9 | N/A | Independent | medium | Mar 15, 2023 | arxiv.org |
| GPT-3.5 Turbo | 41.5 | N/A | Independent | medium | May 28, 2020 | arxiv.org |
| meta-llama/Meta-Llama-3-8B-Instruct | 25.7 | N/A | Independent | medium | Feb 27, 2023 | arxiv.org |
Every number links to where it was published. Low-confidence rows are muted and need per-cell verification.
Reproduced by the community
All runsLoading community runs…
About
MedConceptsQA is a dedicated open source benchmark for medical concepts question answering. The benchmark comprises of questions of various medical concepts across different vocabularies: diagnoses, procedures, and drugs.
Available in EleutherAI lm-evaluation-harness as task family med_concepts_qa (15 tasks).