SK Telecom announced on the 18th that it has greatly enhanced the multi-LLM agent feature by adding Google’s latest LLM model ‘Gemini 2.0 Flash’ to its AI agent ‘A.X,’ along with ‘Perplexity Sonar Pro’ and ‘GPT o3-mini.’
With this update, users of ‘A.X’ can freely utilize 12 global leading AI models, including SKT’s own LLM ‘A.X,’ and the GPT models (4o, 4o mini, o3-mini, o1-mini, o1-preview), Claude (Claude 3.5 Sonnet, 3.5 Haiku, 3 Opus), and Perplexity (Perplexity Sonar, Sonar Pro).
Specifically, the newly added ‘Gemini 2.0 Flash’ is Google’s latest AI model released last February, significantly improving response speed and performance over the previous ‘1.5 Flash.’ In major benchmarks, it recorded twice the speed of the ‘1.5 Pro’ model and also supports real-time search results and source page navigation, allowing users to directly verify the basis of AI-generated responses.
Earlier this month, SKT also added the ‘Perplexity Sonar Pro’ model and ‘GPT o3-mini’ model to ‘A.X’. The ‘Perplexity Sonar Pro’ uses twice as many search sources as the existing models to provide more accurate answers and has improved conversational context understanding. The ‘GPT o3-mini’ excels in interpreting advanced mathematical formulas and generating programming code.
Users can access AI models on PCs through the ‘A.X’ website (https://adot.ai) and use the ‘Compare with Other AIs’ feature to choose their desired AI model for responses. When the same question is applied to multiple AI models, comparisons can be made in one go without copying and pasting, and the crossover conversation feature ensures that the conversation context is naturally continued even when switching AI models during a chat.
Kim Yong-hoon, head of SKT’s A.X business unit, said, “‘A.X’ is the optimal platform for comparing and utilizing core global AI models from ‘A.X’ to ChatGPT in one place,” and added, “We will continue to enhance user experience by adding various AI models and convenience features.”