Google has opened entry to the 2-million-token context window of the Gemini 1.5 Pro AI mannequin. The firm is also giving builders entry to code execution capabilities within the Gemini API and making the Gemma 2 mannequin obtainable in Google AI Studio.
These bulletins have been made on June 27. The 2-million-token context window of Gemini 1.5 Pro, beforehand obtainable through a waitlist, now could be obtainable for all builders.
Google famous that because the context window grows, so does the potential for enter value. To assist builders reduce prices for duties that use the identical tokens throughout a number of prompts, Google has launched context caching within the Gemini API for Gemini 1.5 Pro and Gemini 1.5 Flash.
To unlock capabilities for math and information reasoning for builders, Google additionally has enabled code execution in Gemini 1.5 Pro and Gemini 1.5 Flash. The code execution function permits the mannequin to run Python code and study iteratively from outcomes till a ultimate desired enter is reached. This is taken into account a primary step ahead with code execution as a mannequin functionality. It is obtainable through the Gemini API and in Google AI Studio underneath “advanced settings.”
Google additionally mentioned it has made the Gemma 2 mannequin obtainable in Google AI Studio for experimentation, and it’s working to supply tuning for Gemini 1.5 Flash to all builders. Text tuning in Gemini 1.5 Flash now could be prepared for pink teaming and can be rolled out step by step to all builders. All builders may have entry to Gemini 1.5 Flash tuning by the Gemini API and in Google AI Studio by mid-July, the corporate mentioned.
Copyright © 2024 IDG Communications, Inc.
