Cover Image for Everything you need to know about Kimi K3.1 Pro: A conversation with Moonshot AI and Together AI
Cover Image for Everything you need to know about Kimi K3.1 Pro: A conversation with Moonshot AI and Together AI
Avatar for Together AI Calendar
Hosted By

Everything you need to know about Kimi K3.1 Pro: A conversation with Moonshot AI and Together AI

YouTube
Registration
Welcome! To join the event, please register below.
About Event

​Presented in partnership with the Moonshot AI and Together AI teams.

​About Event

​Join Feihu Tang from the Moonshot AI team and Austin Silveria, Kevin Cui and Zain Hasan from Together AI for a technical session on how the Moonshot team built Kimi K3.1 Pro by post-training Kimi K3.

​Most models can handle a prompt, tool call, or coding task in isolation. Real-world agent workflows need to carry context across long-running sessions, navigate repositories, orchestrate terminal tools, interpret visual feedback, and connect research to executable work. Keeping all of that moving coherently becomes harder as tasks grow longer and more complex.

​Kimi K3.1 Pro is Moonshot AI’s post-trained upgrade to Kimi K3. It’s a mixture-of-experts model (2.8T total parameters, 104B active) with a 1-million-token context window. Moonshot reports a Terminal-Bench 4.0 score of 63.6%, up from 17.2% on base K3. It delivers substantial improvements across coding, 3D creation, financial research, and other agentic capabilities.

​This is a learning-first webinar, ideal for folks building and scaling LLM-powered apps, tools, and agents who need to understand what’s different about K3.1 Pro and how to use it in production.

​​What we’ll cover

  • ​Why K3 needed a post-training pass: where long-horizon coding and agent workflows hit limits on the base model

  • ​What changed in K3.1 Pro: what the post-training work Moonshot did on top of K3’s architecture enables for agentic applications

  • ​What Moonshot learned from building Kimi K3: high-level lessons from developing a model for long-horizon coding, reasoning, and knowledge work

  • ​K3.1 Pro in practice: examples across coding, agents, and the same long-horizon workflows K3 was built for snd how it performs against K3

  • ​Building on Together AI: how to access K3.1 Pro now via the API and run long, tool-heavy agent workloads on Together Serverless Inference.

​Register to join the webinar live, ask questions, and receive a recording of the talk.

​⚠️ Please provide your valid corporate email upon registration.

Avatar for Together AI Calendar
Hosted By