Fireworks AI

🔥 Announcing FireOptimizer/Multi-LoRA 🔥 I didn't expect what I considered to be a small feature launched last year delivered a powerful impact to our customers. I'm excited to announce Multi-LoRA, an important component of FireOptimizer. Personalized experiences are critical to driving greater usage, retention and customer satisfaction for your product. Without Multi-LoRA, deploying hundreds of fine-tuned models on separate GPUs would be prohibitively expensive. With Multi-LoRA, you can now deliver personalized experiences across thousands of users and use cases, without scaling your costs! More specifically, Multi-LoRA has benefits below: -- Fine-tune and serve hundreds of personalized LoRA models at the same cost as a single base model, which is just $0.2/1M tokens for Llama3.1 8B -- 100x cost-efficiency compared to serving 100 fine-tuned models without Multi-LoRA on other platforms with per-GPU pricing -- Convenient deployment on Fireworks Serverless with per-token pricing and competitive inference speeds, or Fireworks On-Demand and Reserved for larger workloads Multi-LoRA is part of FireOptimizer, our adaptation engine designed to customize and enhance AI model performance for your unique use cases and workload. FireOptimizer capabilities include Adaptive Speculative Execution (https://lnkd.in/ejdD-wGG), that enables up to 3x latency improvements, Customizable Quantization (https://lnkd.in/dwpTU233), to precisely balance speed and quality, and LoRA Fine-Tuning (https://lnkd.in/et2UFzDy) to customize and improve model performance. ⚡Cresta uses Multi-LoRA to personalize their Knowledge Assist feature for each individual customer on the Fireworks enterprise platform. "Fireworks' Multi-LoRA capabilities align with Cresta's strategy to deploy custom AI through fine-tuning cutting-edge base models. It helps unleash the potential of AI on private enterprise data." - Tim Shi, Co-Founder and CTO of Cresta ⚡Brainiac Labs helps businesses leverage their proprietary data to fine-tune and deploy models using Multi-LoRA on the Fireworks self-serve platform. “Using Fireworks, clients with limited AI expertise can successfully maintain and improve the solutions I provide. Additionally, students in my course are able to complete real-world fine-tuning projects, dedicating just a few hours per week to the process.” - Scott Kramer, CEO of Brainiac Labs 👉 Read more in our blog post https://lnkd.in/d3_HGRqy

🚀 We've just enabled function calling support for DeepSeek v3 on the Fireworks AI API. Now developers can seamlessly integrate external APIs and real-time data into their LLM applications. With this update, DeepSeek v3 can now: ☑ Fetch real-time data like weather, stocks, and news ☑ Automate workflows and tasks ☑ Interact with external tools and APIs Check out our latest blog post for code examples and implementation details: https://lnkd.in/diGADDWx #AI #LLM #DeepSeek #APIIntegration #DeveloperTools

🚀 Agentic AI Hack Night is TOMORROW! Join us Feb 19 at GitHub SF to: ✅ Build AI agents with DeepSeek AI V3 & hack with R1 ✅ Get $20 in free Fireworks credits to test cutting-edge models ✅ Tackle AI agent challenges & win prizes ✅ Learn from Fireworks AI, CrewAI, Weaviate & VESSL AI in lightning talks ✅ Show off your builds in live demos & connect with top AI engineers 📅 Date: Tomorrow, Feb 19 📍 Location: GitHub SF |⚡ Last chance to register! Don’t miss out! 🔗 https://lnkd.in/e6mPPcit #AI #Hackathon #AgenticAI #LLMOps #GenerativeAI #FireworksAI #DeepSeek #GitHub #VESSLAI #CrewAI #Weaviate

Fireworks AI joins Hugging Face as Official Inference Provider! We're thrilled to announce that Fireworks.ai is now fully integrated into the Hugging Face Hub as a supported Inference Provider. This partnership brings blazing-fast serverless inference capabilities directly to model pages, making it easier than ever to deploy and experiment with state-of-the-art AI models. 🚀 Key Features: - Instant access to models like DeepSeek-R1, Mistral-24B, and Llama-3 - Seamless integration with Hugging Face SDKs - Simple deployment with just a few lines of code - Standard Fireworks API rates with no markup 💡 Pro tip: Hugging Face PRO users get $2 worth of Inference credits monthly! Light up your HF projects with Fireworks AI: https://lnkd.in/dvFxVPMR #TechNews #AI #Partnership

🚀 New SF event: Build Agents with DeepSeek AI 🔥 Agentic AI Hack Night is happening Feb 19 at GitHub SF – are you ready? ✅ $20 in free Fireworks credits – Try DeepSeek V3 for agent workflows & R1 for coding tasks ✅ AI Agent Challenges – Solve real-world problems & optimize workflows ✅ Lightning Talks & Demos – Learn from VESSL AI, CrewAI, Weaviate & Fireworks AI ✅ Prizes for top projects – Show off your build & win big ✅ Networking with top AI builders – Meet devs, engineers & founders 📅 Date: Wednesday, Feb 19 📍 Location: GitHub SF ⚡ Spots are filling up – register now! 🔗 (Link in comments) #FireworksAI #DeepSeek #GitHub #VESSLAI #CrewAI #Weaviate #AIAgents #AIWorkflows #AIHackathon #DeepSeekv3 #DeepSeekR1

Your AI model is only as good as the data it learns from! With MongoDB and Fireworks AI, developers get the best of both worlds: ⚡ Fast, flexible data storage with MongoDB’s document model & vector search. 🎯 Optimized fine-tuning with Fireworks AI’s powerful inference platform. 🔍 Better accuracy for GenAI applications through Retrieval-Augmented Generation (RAG). Learn how this collaboration makes secure, scalable AI development easier than ever in this The Stack article. 👇 https://lnkd.in/e6bzK7mD

The AI shockwave is real yet that doesn't mean there is not clarity forming around what the next 5 years could look like. Chatting to an AI insider like Lin Qiao (Meta, Fireworks AI) lasting themes appear: 1. Open Source Tsunami Comes For AI DeepSeek has (briefly) put Open Source ahead of proprietary AI performance. This is just the beginning of the community movement that will see it win the model wars. 2. Shift To Small Expert Models As LLMs Commoditize Low cost, expert small models unlock the important last mile of value in enterprises and consumer. 3. Enterprises Transform Their Private Data In AI Assets Private data eclipses public internet data, 90% remains untapped, and it will steadily processed to be high value AI assets. Read the full breakdown of each and listen to our conversation in the link below 👇🏼

DeepSeek's latest models—v3 and R1—bring a game-changing approach to efficiency and performance. So to the tech enthusiasts and AI professionals, here’s what you need to know: → DeepSeek MoE (Mixture of Experts): Activates only the most relevant experts per token, ensuring smarter, more efficient AI that scales without breaking the bank. → FP8 Pre-training: Their FP8 pre-training strategy slashes memory usage while maintaining precision—delivering performance without compromise. → Fireworks AI Inferencing: Integrating Fireworks has supercharged the inference efficiency, pushing the boundaries of what’s possible in real-time AI performance. Dive into the full article and join the conversation on the future of scalable AI innovation: https://lnkd.in/eJRBHNHN ----------- Want to run DeepSeek v3 and R1 models for production workloads? Try it on fireworks.ai 🚀

🚀 DeepSeek R1 leads in performance—now even more accessible DeepSeek R1 continues to lead for developers, outperforming competitors in key areas: 1️⃣ #1 Open Source Model for Web Arena – Ahead of O3-mini-high in web-based tasks (https://lnkd.in/d_FWirX4) 2️⃣ Top Choice for Coding – Leading O3-mini with enhanced style control 3️⃣ Fireworks Deployment Optimized for Performance – Now even faster with latency and throughput improvements. Dedicated deployments achieving 95 t/s on Fireworks. Reach out if you are interested in learning more. (https://lnkd.in/dV4UVJcY) Plus, DeepSeek R1 is now even more accessible, with V3 soon to follow: ✅ New pricing for R1: Now $3/$8 (was $8/$8) – effective today ✅ Updated pricing for V3: Moving to $0.75/$3 (from $0.9/$0.9) on Monday, Feb 10 ✅ Flexible input/output pricing based on industry best practices and user feedback Want to see for yourself? Try DeepSeek R1 in the playground today. 🔹 Use R1 now→ https://lnkd.in/g9Xt4grp Got questions? Explore the DeepSeek FAQs. 🔹 Learn more here → https://lnkd.in/d-u4u7pt

Let’s give DeepSeek R1 eyes! A smart reasoning LLM is good, but a smart reasoning VLM is even better! We’re excited to showcase how DeepSeek R1 can now process and reason over both text and images using Fireworks AI Document Inlining. This feature expands DeepSeek R1’s capabilities to multimodal analysis, unlocking new possibilities for AI research and applications. What does this mean? ✅ Smarter research analysis – Extract key insights from papers with text + figures ✅ Better document processing – Handle multimedia content seamlessly ✅ Enhanced AI applications – Build more powerful assistants and knowledge tools Document Inlining is easy to use—just append #𝘁𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺=𝗶𝗻𝗹𝗶𝗻𝗲 to your document URL via our OpenAI-compatible API. Check out our latest blog to see how DeepSeek R1 is getting “eyes” and leveling up multimodal AI! 👇 https://lnkd.in/dZbi9ETD