Google launched the successor of the Gemini 1.5 household of AI fashions, dubbed Gemini 2.0, on Wednesday. The brand new AI fashions include improved capabilities together with native help for picture era and audio era, the corporate highlighted. Presently, the Gemini 2.0 mannequin is out there in beta to pick builders and testers, whereas the Gemini 2.0 Flash AI mannequin has been added to the online and cellular apps of the chatbot for all customers. Google stated the bigger mannequin may also be pushed to its merchandise quickly.
Google Gemini 2.0 AI Fashions
9 months after the discharge of the Gemini 1.5 collection of AI fashions, Google has now launched the upgraded model of the massive language mannequin (LLM). In a blog post, the corporate introduced that it was releasing the primary mannequin within the Gemini 2.0 household — an experimental model of Gemini 2.0 Flash. The Flash mannequin usually accommodates fewer parameters and isn’t match for advanced duties. Nonetheless, it compensates for it with low latency and better effectivity than bigger fashions.
The Mountain View-based tech big highlighted that the Gemini 2.0 Flash now helps multimodal output akin to picture era with textual content and steerable text-to-speech (TTS) multilingual audio. Moreover, the AI mannequin can also be outfitted with agentic capabilities. 2.0 Flash natively calls instruments like Google Search, code execution-related instruments, in addition to third-party capabilities as soon as a consumer defines them by way of the API.
Coming to efficiency, Google shared Gemini 2.0 Flash’s benchmark scores based mostly on inner testing. On the Huge Multitask Language Understanding (MMLU), Natural2Code, MATH, and Graduate-Degree Google-Proof Q&A (GPQA) benchmarks, it outperforms even the Gemini 1.5 Professional mannequin.
Gemini customers can choose the experimental mannequin from the mannequin selector choice positioned on the high left of the online and the highest of the cellular app interface. Aside from that, the AI mannequin can also be obtainable by way of the Gemini utility programming interface (API) in Google AI Studio and Vertex AI. The mannequin can be obtainable to builders with multimodal enter and textual content output. Picture and text-to-speech capabilities are at present solely obtainable to Google’s early-access companions.
Catch the most recent from the Shopper Electronics Present on Devices 360, at our CES 2025 hub.
About The Author
