169Pi

A 9B AI model. Running on an iPhone 17 Pro. No cloud required. What if your AI assistant worked just as well on a flight, in a remote location, or with your internet switched off? We are working to make that a reality with our 9B model. We’ve successfully run it locally on the iPhone 17 Pro, recording 230.65 tokens/sec in prompt processing, 9.59 tokens/sec in generation, and 7 GB peak memory usage in our benchmark. This is what makes local AI interesting beyond the technology itself. Developers can explore apps that work offline. Enterprises can bring AI closer to sensitive workflows. And users can access AI for coding, research, and writing without every interaction depending on a remote server. AI that works without the internet changes where, when, and how you can use it. We are pushing this further at @169Pi_ai. Our work on the model is ongoing, with a focus on making the model more accurate, faster, and more efficient on consumer hardware. Getting all three right on a phone is the challenge we’re working to solve. The iPhone 17 Pro is one step. There’s more to come. Tomorrow, we’ll share how Alpie Seed runs on other devices. The goal is simple: make capable AI available wherever you need it, not just wherever you have a connection.

A 9B model scoring 61.2 on SWE-Bench Verified is interesting. Making that level of coding capability available to run locally is what makes it exciting. At 169Pi, we are working towards AI that can tackle real software engineering and scientific reasoning tasks without always depending on the cloud. Think about what that could mean for developers, enterprises working with proprietary code, researchers handling sensitive work, or anyone who wants capable AI even without an internet connection. Less dependence on remote infrastructure. More control over where intelligence runs. We are pushing the limits of what's possible with efficient models, and we're getting ready to show you what we've built. Starting now, we'll be sharing something new we are building every other day. The research, the experiments, the breakthroughs, and the work behind them. More soon.

It’s here. Alpie is now on iOS and Android. The way we work with AI is changing. We’re building Alpie to make that experience more useful, more collaborative, and more accessible, wherever you are. Now, you can take Alpie with you. Your ideas don’t have to wait until you’re back at your desk. Download #Alpie and experience it for yourself. iOS: Android: We’d love to hear your thoughts. Your feedback will help shape what comes next.

We’re excited to partner with @magtappindia, the Desh Ka AI Browser, bringing 169Pi’s Alpie AI models into a browser with 6M+ users across 106+ countries. MagTapp is now rebuilding the browsing experience around AI and expanding to iOS, taking its journey from visual learning to AI-native browsing. The idea is simple: access to information isn’t enough. Understanding is. Together, we’re bringing AI directly into the way people browse, search, read, translate and learn, with multilingual intelligence designed around how India actually uses the internet. MagTapp’s vision is not an English-first product adapted for India, but an experience built for India from the ground up. Built in India. Powered by Alpie. Designed for the next billion. MagTapp × 169Pi. The browser is becoming intelligent.

The biggest festival of Open Source is just hours away. Community leads and organisers can set up #GoodFirstAlpie as an icebreaker or a side quest for developers during their breaks. Contribute together, get on the leaderboard, and increase your chances of being among the first to land on our GitHub page. Swags are on us. Visit:

Your first open source PR shouldn’t feel like a job interview. #GoodFirstAlpie is live. Pick a starter issue, add your tile to the contributor wall, and Alpie writes your entry prompt right on the page. Hacktoberfest is days away. Warm up with us. Contribute: Discord:

Official Verification — Alpie (PIE) Alpie (PIE) is the utility token of the Alpie ecosystem developed by 169Pi. Official contract: aqQRs3UJmHs7ktoLo46HZAzvG9P3oCoZ41nENqFbory Solana: This post is being published as official verification for CoinMarketCap listing request #1459878. Official website:

Open source starts with a first contribution #GoodFirstAlpie is live. Come make your first contribution to @169Pi_ai, get Hacktoberfest-ready, and build something with us. 🚀 Let’s build together.

Keeping my promise - #GoodFirstAlpie is up! Your chance to be on the landing page of @169Pi_ai, make your first contribution to open-source, get ready for @hacktoberfest and earn exclusive swags. contribute today:

Frontier intelligence. Smaller footprint. Run on your devices.

Alpie Seed Coming Soon!

We are selecting 10 developers for early access to Alpie Seed 🌱 Want to be one of the first to try this SOTA 4-bit multimodal model? Comment below, and you might get selected.

AI is better when you don't have to work alone. With Alpie, you can invite your colleagues or peers into a shared workspace and start researching, building, and solving problems together in real time. Assign roles, work across chats, and move from an idea to something useful without switching between a dozen tools. The interesting part is how collaboration and personalisation work together. Alpie's memory is designed to keep your personal conversations separate from collaborative chats, so your AI experience becomes more relevant to you without mixing in everything your team is working on. Your AI. Your workspace. Your team. Built for people who want to do more, together. Try Alpie:

For a while, progress in AI was largely framed around one question: how much bigger and more capable can we make the model? The next phase may be a different optimisation problem: how much capability can we deliver per bit, per watt, and per gigabyte of memory? That’s why low-precision inference, quantization, distillation and hardware-aware model design are becoming increasingly interesting. Not because smaller is inherently better, but because efficiency changes where intelligence can actually exist. A capable model that can run on more hardware, closer to the user, and without a massive infrastructure footprint opens up an entirely different set of possibilities, and multimodal models make this problem even more interesting. Something new is coming this month.

Excited to be partnering with Avocado AI! 🥑🤝 We’re looking forward to building a long-term relationship and powering faster, smarter, and better responses for users across the Avocado AI ecosystem. Just getting started — excited for what’s ahead!

We’re excited to announce @169Pi_ai as Avocado AI’s official LLM provider. Their Alpie suite of models will power key capabilities across Flows and Super Agent, helping customers create high-quality ads at lightning speed. This partnership will be especially impactful in Flow Canvas and Director Mode(which soon will be released to general public and out of closed beta). It’s another step toward empowering brands and creators across both Web2 and Web3.

Guest Mode was already part of Alpie. Today, it gets a much better experience. A cleaner interface. A more intentional workflow. The ability to save conversations only when you decide they're worth keeping. We think privacy and flexibility should feel effortless, and not like settings you have to manage. Try it now on Alpie!

The industry often measures progress by how much larger models become. We are asking a different question. How much capability can we preserve while making intelligence dramatically easier to deploy? Over the last few months, we've been training and benchmarking a new 9B 4-bit multimodal model built around that idea. Coding. Software engineering. Scientific reasoning. Multimodal understanding. Efficient enough to run where larger models simply can't, but delivering frontier performance. We think deployability deserves as much innovation as intelligence itself. Looking forward to sharing what we've been building.

Most work doesn't happen inside a single file. A financial model is paired with a report. A spreadsheet references a presentation. A PDF explains the numbers. Screenshots capture the missing context. Alpie now supports spreadsheets/excel alongside your other attachments, allowing you to reason across various files, PDFs, images, documents, and more within the same conversation. AI becomes far more useful when it understands the complete picture. Try it now!

Why are we still treating AI like it can only do one thing at a time? While one conversation is researching, another should be building your presentation. While that's happening, another can review your documents, generate a website, or analyse your spreadsheet. That's exactly how Alpie works now. Run up to 5 conversations simultaneously and switch between them freely. We'll notify you when each task is complete. AI shouldn't make you wait. It should work with you. The future of AI isn't one conversation. It's parallel execution. Early release. We'll be shipping improvements rapidly over the coming days. Try it, push it to its limits, and tell us where we can make it better.

Alpie now has Incognito Mode. your fears. your ambitions. your space. no history. no judgment. just answers. and if the answer’s too good to lose? tap Save Session – only you decide what stays. try free →

Today, we are excited to introduce the highly requested Incognito Mode on Alpie. This feature creates a space for conversations that don’t need to be remembered. By default, sessions are private and are not saved to your history. However, sometimes unexpected insights emerge: - A rough idea can develop into a valuable insight. - A quick question may lead to important research. - A temporary conversation could be worth keeping. That’s why you have the option to save a session whenever you choose. It won’t be done automatically or before you decide. Incognito Mode is available now. We look forward to hearing your feedback on this, as it's still a fresh feature!

@Magtapp × @169Pi_ai 🤝
We're bringing Alpie AI into MagTapp. 6M+ downloads, users in 106 countries, now on iOS. Search with direct answers is live, with summaries, translation and visual explanations coming next. 
Built in India, for Bharat and the world.