https://datafranca.org/wiki/index.php?action=history&feed=atom&title=InternVLInternVL - Historique des versions2026-04-09T17:46:05ZHistorique des versions pour cette page sur le wikiMediaWiki 1.39.5https://datafranca.org/wiki/index.php?title=InternVL&diff=117036&oldid=prevPitpitt : Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' InternVL''' == Anglais == ''' InternVL''' A new family of open-source multimodal large language models that significantly advances capabilities in versatility, reasoning, and efficiency. The models range from 1B to 241B parameters and achieve state-of-the-art performance among open-source models while narrowing the gap with commercial systems like GPT-5. InternVL3.5 achieves impressive per... »2025-09-20T14:10:02Z<p>Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' InternVL''' == Anglais == ''' InternVL''' A new family of open-source multimodal large language models that significantly advances capabilities in versatility, reasoning, and efficiency. The models range from 1B to 241B parameters and achieve state-of-the-art performance among open-source models while narrowing the gap with commercial systems like GPT-5. InternVL3.5 achieves impressive per... »</p>
<p><b>Nouvelle page</b></p><div>==en construction==<br />
<br />
== Définition ==<br />
XXXXXXXXX<br />
<br />
== Français ==<br />
''' InternVL'''<br />
<br />
== Anglais ==<br />
''' InternVL'''<br />
<br />
A new family of open-source multimodal large language models that significantly advances capabilities in versatility, reasoning, and efficiency. The models range from 1B to 241B parameters and achieve state-of-the-art performance among open-source models while narrowing the gap with commercial systems like GPT-5.<br />
InternVL3.5 achieves impressive performance across multiple benchmarks. The largest model, InternVL3.5-241B-A28B, attains state-of-the-art results among open-source models and narrows the performance gap with GPT-5 to just 3.9% on general multimodal tasks. On reasoning benchmarks, the models show substantial improvements, with InternVL3.5-8B achieving 73.4 on MMMU and InternVL3.5-241B-A28B reaching 77.7. The Cascade RL framework provides up to 16.0% improvement in overall reasoning performance compared to the predecessor InternVL3.<br />
<br />
<br />
== Source ==<br />
<br />
[https://huggingface.co/papers/2508.18265 Source : huggingface]<br />
<br />
[[Catégorie:vocabulary]]</div>Pitpitt