Gemini 1.5 Pro now comes with 2M context tokens. For context, GPT4 has 128K. What it means: 1️⃣ You can now put more data into your context window thereby increasing your output accuracy, all else being equal 2️⃣ At a text input price of USD0.70/1M tokens, this means a single prompt on Gemini 1.5 Pro could set you back by up to USD1.40, not including any output response Is there a way to optimise accuracy and cost? One way is to use context caching offered by Gemini (but not available on other LLM, such as GPT4o), which may reduce the price to pay under certain conditions (not all, yeah, it’s a bit complicated). Another way is to use RAG, which is applicable to all LLM, but its implementation is highly skill-dependent. Finally, and probably the easiest way, is to engage a professional and watch them do the magic, because a GenAI solution, when well-designed and well-developed, works like magic.
shin wee chuang 庄欣伟’s Post
More Relevant Posts
-
Google has announced the availability of a 2 million token context window for their Gemini 1.5 Pro AI model - enabling the model to work on even longer inp t - and additionally the availability of smaller Gemma 2 models https://lnkd.in/e_fwNYrD. #google #gemini #gemma2 #generativeai #largelanguagemodel #artificialintelligence
Gemini 1.5 Pro 2M context window, code execution capabilities, and Gemma 2 are available today
developers.googleblog.com
-
Exciting updates for developers! Check out the new features for the Gemini API and Google AI Studio, designed to enhance your AI projects and streamline development. Unlock powerful tools and capabilities to take your applications to the next level! Read more: https://bit.ly/4bAVyjc #GoogleAI #GeminiAPI #AIDevelopment #DeveloperTools
Gemini 1.5 Pro 2M context window, code execution capabilities, and Gemma 2 are available today
developers.googleblog.com
-
Tool of the day AirOps: on Ai Jumble-Building a thriving AI community. Unlock Growth with AI-Powered Workflows! 🚀 Design and expand integrated workflows using cutting-edge AI models like GPT-4-turbo, DALLE-3, and Claude 2. Start swiftly with customizable templates, fine-tune your prompts, and experiment with diverse models from OpenAI, Anthropic, and Meta to elevate your solutions. Build, test, and launch your v1 solution in under 10 minutes—combine prompts, data, code, and APIs visually for the perfect outcome. Elevate your AI's capabilities with managed Memory Stores, effortlessly setting up and maintaining your document libraries and data connections. Elevate your AirOps Apps to the next level! https://bit.ly/4abbHMO Join Ai Jumble-Building a thriving AI community,-for more exciting #ai tools #AIWorkflows #IntegratedAI #GrowthHacking #AISolutions
AirOps | Create, test, deploy and scale AI Apps with AirOps Studio.
airops.com
-
Head of Artificial Intelligence at Squaredev | MSc Applied Economics & Data Analysis | Patras AI meetup co-organizer | Medium writer
I have some exiting news to share! 🎉 Following my recent release of the 'streamlit-token-craft' library, as I promised I share my latest contribution: a comprehensive tutorial now live on my Medium account in Towards AI! Read the full article here: https://lnkd.in/dxF7-kTz Stay tuned for more updates, and let's keep pushing the boundaries of open-source development! #streamlit #opensource #customcomponent #development
Brand New Streamlit Component: Token Key Mastery
pub.towardsai.net
-
Independent web performance consultant. Google Developer Expert & Microsoft MVP (reconnect)
I just published a free Gemini AI web course. Check out the videos and learn how to build a predictive support functionality! https://lnkd.in/guheB4Y4 #javascript #ai
-
-
🏗 In a nutshell, this article teaches you how to use Vercel's V0.dev clone project to level up your AI skills. From using Next.js to CopilotKit, this tutorial is like training wheels for an AI motorcycle - start your engines! 💌 Subscribe to get fresh news every day: https://0xcafe.news Author: Ekemini Samuel
How to build: a v0.dev clone (Next.js, GPT4 & CopilotKit)
dev.to
-
Meet the new AI test writing tool for developers: CodiumAI In this article from The New Stack, David Eastman dives into a platform that has ‘Generating meaningful tests for busy devs’ as its homepage elevator pitch – read on to see what he made of it: https://lnkd.in/eGmPUqk3 #AI #SoftwareTesting #Testing #SoftwareDevelopers
Make Your Dev Life Easier by Generating Tests with CodiumAI
https://thenewstack.io
-
Classification in Machine Learning Introduction to Machine Learnin... #ai #javascript #machinelearning #prodsenslive #Software #webdev https://lnkd.in/dbm-GMNC https://lnkd.in/dMz79_hi
Classification in Machine Learning - ProdSens.live
https://prodsens.live
-
We have updated our documentation for Onvo AI after hearing your feedback, making it much easier to find what you are looking for. It has a quickstart guide to get you up to speed in no time, API references with an interactive playground and documentation for all our libraries. You can read more about it at https://lnkd.in/g6iKAX-d You can go straight to the documentation at https://docs.onvo.ai Looking forward to your feedback! #documentation #developer #ai #dashboards #analytics
Launch update: Updated documentation and libraries
onvo.ai
-
In the world of AI, OpenAI's GPTs have been a game-changer, especially in simplifying the use of RAG (Retrieval-Augmented generation) systems. Prior to the advent of GPTs, deploying and maintaining RAG systems was a relativelly complex task, reserved for those with deep technical expertise. My latest article sheds light on this transformation. I take you through the journey of building a basic RAG system and then comparing it with a custom GPT. This exploration reveals how GPTs are revolutionizing the way we design systems to deliver tangible business value. https://lnkd.in/eDfDaDDB #GPTs #RAG #generativeai #ai #artificialintelligence
Implementing RAG with Vue3, FastAPI and Qdrant and how is this compared to OpenAI GPT’s
medium.com
