Just wrapped an incredible few days at NeurIPS in San Diego, and we left more energized than ever. The density of brilliant researchers, founders, and builders all in one place was unreal. Every hallway conversation, poster session, and booth stop turned into a deep dive on something fascinating. I met teams pushing boundaries in agentic systems, multimodality, infrastructure, and everything in between. The vibe was electric: so much curiosity, innovation, and genuine excitement about where AI is heading. I walked away inspired, full of ideas, and grateful for the chance to learn from so many sharp minds. If we connected this week, thank you. If we missed each other, let’s fix that. Until next year, NeurIPS PS - San Diego weather was amazing! #neurips #ai
Deep Infra Inc.
Technology, Information and Internet
Palo Alto, California 1,931 followers
Fast ML inference. Run top AI models using a simple API.
About us
Let Deep Infra run your ML infrastructure. Just use our top AI models using a simple API or deploy your own model with us.
- Website
-
https://deepinfra.com
External link for Deep Infra Inc.
- Industry
- Technology, Information and Internet
- Company size
- 2-10 employees
- Headquarters
- Palo Alto, California
- Type
- Privately Held
- Founded
- 2022
Locations
-
Primary
Get directions
Palo Alto, California 94306, US
Employees at Deep Infra Inc.
Updates
-
We’ll be at NeurIPS in San Diego this week - excited to connect, share what we’re up to, and see what you’re building. If you’re around, let’s link up.
-
Check out the Pruna AI image generation and editing models live on Deep Infra. You can do it all in 1 sec which is quite impressive without compromising quality. Check it out here https://lnkd.in/gjsAUpdw #ai #inference
Pruna AI just leveled up. We’ve officially moved from an “Optimization Framework” to a "AI Model Lab" for Performance Models, and, today, we’re shipping our first generation of INSTANT SoTA Image Models: 📷 P-Image 📸 P-Image-Edit We went after the three pain points everyone quietly complains about: • true seed diversity • bias-resistant generation • sharp fine-detail accuracy And… yeah, we cracked it 🤯 📈 Numbers (because opinions are cute, but benchmarks matter): • P-Image → 0.5s inference, best quality in its category • P-Image-Edit → 0.9s inference, redefining what “instant editing” even means 💸 Pricing: • P-Image → $0.005/output • P-Image-Edit → $0.01/output That’s the cheapest, fastest, highest-quality 1K models on the market. Full stop. We built these models to fill the biggest creative gaps we kept hearing from users. They’re filled now. And yes... we’re absolutely not stopping here! Thanks to our early testers and launch partners: Replicate(Andreas, Luis D.) - Prodia (Mikhail , Monty) - Runpod (Hailong, Zachary) - Runware (Ioana, Flaviu) - Deep Infra Inc. (Leily, Oguz) - Wiro AI (Alican) - Segmind (Rohit, Shrey) And special kudos to ALL the Pruners who did a fantastic job! 🩷
-
-
We have launched the new FIBO model from Bria AI It is #1 in alignment, #1 in aesthetics, and #1 smallest model compared to all other open source models. Remember you can use any Bria models for free till the end of the year, so take advantage while you still can. https://lnkd.in/gXKRx5Y5 #imagemodel #inference #ai
-
-
We’re at NVIDIA GTC Washington DC next week (Tue–Wed)! Who’s going? Come say hi at Booth I-9, we'd love to meet you. Reply if you want to sync IRL 🔋🤝 #Nvidia #GTC
-
If you did not know we have the Google Veo 3.1 model on Deep Infra. Check out the video our production team made. Hollywood may need to watch out 😆 #inference #ai #Veo3.1
-
Come join us for trivia night in SF on Oct 22 with NVIDIA and vLLM. All things open-source and inference. RSVP: https://luma.com/cpgzpcwt
NVIDIA + Open Source AI Week 2025 – powered by partnerships, events, and community 👏 We’re excited to bring together technology, community, and collaboration at Open Source AI Week 2025, hosted by The Linux Foundation. From informal meetups to hackathons, panels to poster sessions, here’s the scoop on where you can join us in to advance open-source AI 👇 🥤 AI Dev Night with Unsloth AI & Mistral AI — Join us for boba and talks on training & deployment with RTX AI PCs. 🧩 Trivia & Community with Deep Infra Inc. & vLLM — A fun, interactive quiz night to connect engineers, practitioners, and open-source devs. 🧑💻 GPU MODE IRL Hackathon — Over 215 networked NVIDIA B200 GPUs, we are joining other mentors from Thinking Machines, Unsloth AI, PyTorch, Periodic Labs, Mobius Labs, Google DeepMind, and more — courtesy of Nebius. 🙌 PyTorch Conference 2025 — NVIDIA-led sessions, posters, meetups, and panels aligned with the flagship event. 🤖 Open Agent Summit — NVIDIA Developer Advocate Mitesh Patel will join a panel on The Future of Agents & Human-Agent Collaboration. 🧠 Measuring Intelligence Summit — Vivienne Zhang (Senior PM, Generative AI Software) will speak on reasoning models, benchmarks, and superintelligence 🤓 Technical Sessions & Posters — Covering topics like Lightweight, High-Performance FSDP on NVIDIA GPU, Scaling KV Caches for LLMs, and more. ⚡ Dynamo & Dine with Baseten — hands-on LLM inference & scaling 💬 Model Builders Meetup with NVIDIA Nemotron & Prime Intellect — open frontier models + RL 🔗 Stay in the loop — bookmark our event page for updates as we add more → https://nvda.ws/48ExDSb
-
-
Now Factory users get 25% off the first $20 of usage every month. It’s applied automatically with BYOK. Add your DeepInfra API key and send your requests.
Starting today, you can use any open-source model to power your Droids. Droids achieve the highest scores across all open-source models on Terminal-Bench. We find GLM 4.6 to be the most performant, remarkably achieving a score in Droid that beats Sonnet 4 in Claude Code. In less than 1 minute, Bring Your Own Keys (BYOK) into the Factory CLI (free tier) to run GLM 4.6, Qwen3 Coder, DeepSeek V3.1, and more Run your inference with some of our launch partners Baseten OpenRouter Fireworks AI Deep Infra Inc. Hugging Face Ollama. BYOK setup docs below.
-
-
Think you can out-nerd your peers? Join NVIDIA, Deep Infra Inc. , and vLLM for trivia night in San Francisco. From GPUs to pop culture, we’ll see who’s got the range. 🚩 7 Social SF | Tuesday, October 22 | 6:00–9:00 PM RSVP here: https://lnkd.in/gAC4azbq #AIMeetup #AITrivia #DeepInfra
-
We are teaming up with NVIDIA again for a casual AI meetup in San Francisco on October 22. Trivia, beer, and good conversation with the AI community. Free to attend, but spots are limited. 7 Social SF | Tuesday, October 22 | 6:00–9:00 PM Register here: https://luma.com/cpgzpcwt #AICommunity #MeetupSF #TriviaNight
This content isn’t available here
Access this content and more in the LinkedIn app