Sharing is caring!

xLAM models, particularly the 7B and 1B versions, demonstrate superior function-calling capabilities, achieving significant ranks on the Berkeley Function-Calling Leaderboard, surpassing even advanced models like GPT-4 and Claude-3. The effectiveness of xLAM models, trained with the APIGen datasets, is evidenced by their high rankings on the Berkeley Function-Calling Benchmark (BFCL). The xLAM-7B model ranks 6th, outperforming multiple versions of GPT-4 and other notable models such as Llama3 and Gemini. The smaller xLAM-1B model also performs exceptionally well, securing the 24th position and surpassing models like GPT-3.5-Turbo and Claude-3 Haiku. These results highlight the efficacy of the APIGen pipeline in producing robust datasets that significantly enhance function-calling models' performance. The APIGen framework's ability to filter out low-quality data ensures the high reliability of the generated datasets, making xLAM a powerful tool in the realm of functional API execution.

Visit
Find us on AI Scores

Sharing is caring!