GPT-OSS-120B agent with 9B model hits 64.0% on BrowseComp+
Original titleIn agentic search, a GPT-OSS-120B agent using the 9B model answers 64.0% of BrowseComp+ questions correctly, 4.9 points ahead of the next...
AISummary
In agentic search, a GPT-OSS-120B agent using the 9B model answers 64.0% of BrowseComp+ questions correctly, 4.9 points ahead of the next ColBERT model. It also uses fewer searches than any baseline.
Source: Perplexity · x.comPublished · added here