The competition between open-source and closed-source AI has intensified following OpenAI's announcement of its Deep Research initiative. Hugging Face quickly responded with its own version, achieving 55% accuracy in comparison to OpenAI's 67%. Although Hugging Face's model ranks highest among open submissions, it admitted that full parity with OpenAI's advanced o3 model remains a work in progress. The landscape illustrates a significant shift where open-source AI is striving to match, and eventually surpass, established closed-source systems, highlighting broader implications for AI transparency and innovation.
OpenAI's Deep Research achieved 67% accuracy on the GAIA benchmark, while Hugging Face's open counterpart reached 55%, marking significant competition in AI development.
Hugging Face acknowledged that their open Deep Research has limitations in competition with OpenAI's offering, citing the need for expanded range in browser interaction.
OpenAI's version, built on the o3 model, delivered promising results in new benchmark tests, highlighting the advancements in their AI agent capabilities.
The open-source community is rallying in response to OpenAI's innovations, emphasizing the importance of openness and reproducibility in AI research and development.
Collection
[
|
...
]