Update README.md
Browse files
README.md
CHANGED
|
@@ -21,11 +21,11 @@ We build clinically grounded, auditable AI for medicine β co-designed and eval
|
|
| 21 |
|
| 22 |
---
|
| 23 |
|
| 24 |
-
##
|
| 25 |
|
| 26 |
-
Our latest release, **Fully Open
|
| 27 |
|
| 28 |
-
**Apertus-70B-MeditronFO** establishes a new state of the art among fully open medical LLMs, and **Gemma-3-27B-MeditronFO** outperforms MedGemma on HealthBench Hard despite being trained from a fully open pipeline.
|
| 29 |
|
| 30 |
π Browse the [**MeditronFO Collection**](https://huggingface.co/collections/EPFLiGHT/meditronfo) β six models across three families (Apertus, OLMo-2, EuroLLM, Gemma-3), plus the [Fully Open Meditron Corpus](https://huggingface.co/datasets/EPFLiGHT/fully-open-meditron) (~601k clinician-audited examples).
|
| 31 |
|
|
|
|
| 21 |
|
| 22 |
---
|
| 23 |
|
| 24 |
+
## MeditronFO
|
| 25 |
|
| 26 |
+
Our latest release, **MeditronFO** (Fully Open), is the **first** fully open medical specialist LLM familly: open weights, open data, open training recipe, and clinician-vetted corpus construction.
|
| 27 |
|
| 28 |
+
**Apertus-70B-MeditronFO** establishes a new state of the art among fully open medical LLMs, and **Gemma-3-27B-MeditronFO** outperforms MedGemma on HealthBench Hard and AutoMOOVE despite being trained from a fully open pipeline.
|
| 29 |
|
| 30 |
π Browse the [**MeditronFO Collection**](https://huggingface.co/collections/EPFLiGHT/meditronfo) β six models across three families (Apertus, OLMo-2, EuroLLM, Gemma-3), plus the [Fully Open Meditron Corpus](https://huggingface.co/datasets/EPFLiGHT/fully-open-meditron) (~601k clinician-audited examples).
|
| 31 |
|