Introducing Novi!
Keep your harness. Run it on Novita.
Meet Novi: launch Claude Code, Codex, Pi or DeepSeek Harness on Novita models with one command.
Add your API key once. Novi handles the configuration.
More integrations on the way.
#Novi #CodingAgents
It’s free😋
🚀 JEV-27B-VL from @AutoTrustAI is now live on Novita AI!
Launching as a partner ⚡
🎁 Free for 14 days.
A 27B vision-language model for decisions over text and images.
• Yes/no, ratings, or multiple-choice decisions
• Probabilities for each option
• Optional reasoning for harder decisions
New version!
🚀 DeepSeek V4.1 Flash 1017 is now live on Novita AI!
The latest DeepSeek V4.1 Flash update, with improved reasoning, coding, agentic tasks, and visual understanding, plus shorter outputs and faster task completion.
🔹 1M-token context window
🔹 Text and image input
🔹 Tool calling and thinking mode
🔹 OpenAI-compatible serverless API
It's still buggy😅 -> novita.ai/connect
🌉 SF Tech Week Recap: Open Source AI Stack
On Oct 7, builders, researchers, and founders came together in San Francisco for “Open Source AI Stack: Models, Inference, and Agent Harness” as part of #SFTechWeek, with 540+ registered. 🧡
Mao Cheng(@MCheng89333 ), core contributor to Miles and MTS at @radixark, joined Sam Herring (@NousResearch), Walter Korman (@vercel), and Akshay (@ArtificialAnlys) on a panel about how the AI stack is expanding beyond the model, covering inference, post-training, benchmarking, and agent harnesses.
Thanks to @novita_labs for hosting, and SF, thank you for the energy. See you at the next one! 👇
luma.com/SGLang
#SFTechWeek @Techweek_
Alex retweeted
We had 400+ attendees join us at AWS Builder Loft, making this one of the top 3 most-attended events held there this year.
Huge thanks to our speakers, exhibitors, partners, and everyone who came out to spend the evening talking about open models, inference, agents, and the infrastructure behind the next generation of AI products.
Special thanks to @ArtificialAnlys @NousResearch @radixark @vercel @CommandCodeAI @kylon_ai @MiniMax_AI @Tiny_Fish @zeddotdev and @ZooWorkAI!
The best part was seeing just how strong the appetite is for deeper, more technical conversations around the open AI stack.
Appreciate everyone who helped make this happen!
Alex retweeted
SF Tech Week Panel | Speaker Highlight
Excited to welcome Sam Herring (@yaboilyrical), Member of Technical Staff at @NousResearch, to our panel:
Open Source AI Stack: Models, Inference, and Agent Harness
Based on his work with agents and post-training, Sam will:
• Discuss how far open-source models have come since the early days of Nous Research
• Examine whether the next wave of AI innovation is shifting from model training to inference
• Unpack the tradeoffs between Hermes and emerging personal agents
Join us on Oct 7 at AWS Builder Loft in downtown SF.
RSVP → luma.com/novita-oas6
🤖 Made with AI
Opus5.5+skill from @op7418 . One shot for promo video even including music, is it legit?
Alex retweeted
Announcing the speakers for our SF Tech Week panel (10/7), exploring the open-source technologies shaping how the next generation of AI products is built and scaled!
-Walter Korman (@shaper), Tech Lead at @vercel
-Sam Herring (@yaboilyrical), ML Engineering Lead at @NousResearch
-Mao Cheng, Member of Technical Staff at @radixark
-Akshay Deshpande, Member of Technical Staff at @ArtificialAnlys
🤓
Replying to @novita_labs
@novita_labs is SO underrated. They've served a quarter TRILLION tokens for us the last 7 days alone.
And top tier service. We love our partners @concentrateai
Let’s go!
We're teaming up with @a16z for SF @Techweek_!
Join us for a moderated conversation featuring leaders from @ArtificialAnlys, @NousResearch (Hermes Agent), and @radixark. Together, they’ll explore the open-source technologies shaping how the next generation of AI products is built and scaled.
This event is on October 7th at AWS Builder Loft in downtown San Francisco. Comment below to RSVP.
Alex retweeted
Thanks to @ArtificialAnlys for updating its Claude Code + GLM‑5.2 results for Novita.
The updated chart shows approximately 5.2M cached input tokens out of 5.6M total tokens—a materially different observed cache profile from the earlier results.
AI moves too fast. Half the day I don't even know what shipped.
So I built Frontier — my first open-source project. It aggregates today's AI news with sources, and ships as an Agent Skill so your agent can pull the latest and trace every claim.
Live site : frontiermemo.com
Github : github.com/Alex-yang00/front…
What's next?
🎥 Launch video ?
📡 More sources ?
🎙️ AI-generated 5-min daily podcast ?
📺 5-min daily video digest ?
Open source, so PRs are welcome.
























