Novita AI
@novita_labs
AI & Agent Cloud for Developers
⚡Join our community:
415 Following 5.7K Followers
Super excited to unveil Hack Week, a new community-led, builder focused track happening for the first time across SF and LA @Techweek_ !
We’re bringing together 150+ technical events across both cities, including 69 hackathons alongside and tons of product demos, workshops and builder meetups.
Events are hosted by Anthropic, Google, Amazon, MongoDB, Mistral, OpenRouter, Datadog, Scale, Exa, Cloudflare, Stanford, Berkeley, USC, and many more.
We’ll culminate the week with SpeedHacks, the official Speedrun Hackathon, featuring tracks hosted by speedrun portfolio companies. The winning team will receive a final-round interview for Alpha or Speedrun.
Show more
Excited to power isolated execution for Dot’s Agent Chat with Novita Sandbox 🙌
We’ve integrated an isolated sandbox infrastructure environment provided by our official partners, @novita_labs, backed by our existing zero-training-data contract.
This is part of our broader initiative to make the Dot platform more agentic by design.
Use Agent Chat to research, analyze and create with AI that gets work done.
- Search the web and turn findings into reports.
- Turn your files into interactive dashboards and visualizations.
- Create PDFs, presentations and custom tools.
- Preview, refine and keep working in the same chat.
- Close the tab. Your task keeps running.
Start free: 3 prompts a day for your first 48 hours.
Show more
Great work by the vLLM team! Happy to have supported these optimization efforts with compute. 🚀
Kimi K3 serving in vLLM now delivers 2.2–2.8x throughput on our B300 benchmark vs v0.27.1.
We break down the work across scheduling, KDA state handling, and MoE kernels, with benchmarks and commands to reproduce the results.
Thanks to the vLLM community for pushing Kimi K3 performance forward!
Read the deep dive:
Show more
Congrats to our partner Command Code on launching their desktop app! 🎉
You can also use Novita models via BYOK.
Give it a try!
Introducing Command Code Desktop App! ⌘
Available today in beta for Mac, Linux, & Windows.
Building the most powerful agentic app, starting with code. Multiple agents, files, Git, terminal, browser previews, plans, and instant edits with /design.
This is just the beginning.
Show more
Open source will win!
🚀 vLLM's Humming backend can run Chord, @novita_labs' open-source W4A16 MoE kernels for Kimi K2.x.
Kernel gains reach 1.33x on H200 TP8 vs tuned public Humming and 2.15x on B300 EP8 decode vs its untuned default.
The indexed path works on compatible vLLM revisions; grouped integration is WIP. Great to see the kernels and benchmarks open-sourced!
Details:
Show more
@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
Real usage is the best signal. ⚡
Ling 3.0 Flash VL is already seeing usage across Kilo Code, Hermes Agent, Claude Code, OpenClaw, Cline, and more.
Free access is still available—give it a try while it lasts.
Show more
Ling 3.0 Flash VL is live on OpenRouter, served by @novita_labs.
A 124B MoE with 5.5B active params, now with native image and video input plus visual agent capabilities. Hybrid instant and reasoning modes, with tool calling.
Show more
Glad to have you building with Novita!
Few days back I was about to get nahcrof credits, about $120, glad my credit card did not work and I went with @novita_labs
Proud to support DeepSeek V4.1 Flash at this scale.
Customers have used 100B DeepSeek V4.1 Flash tokens from @novita_labs over the last 24 hours at an avg of 213 T/S. ⚡️ fast!
Try here:
Which one feels closest to the real thing?
If you had to choose one model to build with, which one would you pick?
One game concept. Four frontier models. One Temple Run-inspired 3D endless runner.
GPT-6 Astra — 4,033 tokens, $1.280423
GLM-5.3 — 4,337 tokens, $0.018591
DeepSeek V4.1 Flash — 7,356 tokens, $0.008181
Kimi K3 — 5,756 tokens, $0.523840
Powered by Novita AI — 200+ models, one API.
Show more
Proud to serve Ling 3.0 Flash VL. ⚡
Native image and video input, hybrid reasoning, and tool calling—ready for the next generation of multimodal agents.
Ling 3.0 Flash VL is live on OpenRouter, served by @novita_labs.
A 124B MoE with 5.5B active params, now with native image and video input plus visual agent capabilities. Hybrid instant and reasoning modes, with tool calling.
Show more
Ling-3.0-Flash-VL is live on @OpenRouter we’re proud to support its Day 0 launch.
Free for two weeks. No local setup. Just call the API and start building multimodal apps and agent workflows.
Ling-3.0-flash-VL is now available on @OpenRouter with two weeks of free access.
No local deployment is required. Call the API and start building multimodal applications and Agent workflows.
OpenRouter:
A special thank-you to @novita_labs for supporting this launch.
Show more
DeepSeek V4.1 Flash is available in HuggingChat 🐋
(absolute must try if you have a HF account!)
🤗 Novita now supports DeepSeek-V4.1-Flash on @huggingface.
• 552B backbone parameters
• Native image and text input
• Up to a 1M-token context window
• Continuously controllable reasoning effort
Show more
Thanks to our day0 partner @novita_labs for the continued support, this time with the "eyes" of the Ling flash model series. Enjoy Ling-3.0-flash-VL for a limited free period! Be creative and enjoy! 🫶
Show more
Excited to partner with Command Code and offer Ling 3.0 Flash Sante for free!
Ling 3.0 Flash Sante is now free in Command Code.
5.1B active params. all subs. all plans
npm i -g command-code@latest
🐐
Start building and see what it can do.
In French we use the word Sante, meaning "to your health", typically used in a toast.
We wish you all enjoying the ling-3.0-flash-sante model, now available with our day0 partner @novita_labs ~ 😃
Good news: the free trial for Ling-3.0-Flash-Sante has been extended to 30 days 🎁
🚀 Ling-3.0-Flash-Sante is now available via Novita on @OpenRouter.
🎁 Free for 14 days.
• 124B total parameters · 5.1B active parameters
• Built for medical reasoning, clinical safety, and long-horizon tasks
Show more
Sante is now available on @OpenRouter and @vercel_dev. Healthcare professionals, researchers, and developers can try Sante free for one month through the OpenRouter API and explore its capabilities on real-world healthcare tasks.
Try it here:
OpenRouter:
Vercel:
A special thank-you to @novita_labs for supporting this launch.
Show more
🚀 Ling-3.0-Flash-Sante is now available via Novita on @OpenRouter.
🎁 Free for 14 days.
• 124B total parameters · 5.1B active parameters
• Built for medical reasoning, clinical safety, and long-horizon tasks
Show more
