From Open Models to Live, Interactive Storytelling - MiniMax x SGLang Summit
Hosted by MiniMax
Register with organizer
What to look forward to
MiniMax, fal, SGLang, and Pickford will show how open models, post-training, and optimized inference power audience-driven storytelling, using Pickford’s procedural murder mystery as a case study. A 20-minute live demo will let audience choices shape the story, followed by Q&A with teams specializing in multimodal models, video deployment, inference, and interactive narratives.
Inside the event
What happens when the audience can change a story as it unfolds? 🎬What happens when the audience can change a story as it unfolds? 🎬
Join MiniMax, fal, SGLang, and Pickford for a behind-the-scenes look at how open models power live, audience-driven storytelling—and help shape an interactive mystery yourself.
Using Pickford’s procedural murder investigation as a shared case study, we’ll follow an audience request through the full stack: from MiniMax H3’s base-model capabilities to fal’s post-training and deployment, SGLang’s diffusion inference, and the story engine that keeps the mystery coherent.
Want a character to wear a party hat? That change once required an asset built in advance. Generative video opens up new possibilities, letting the world respond to unexpected requests while the story keeps moving.
🔍 Inside the session
- MiniMax · Inside H3: The capabilities behind expressive characters, movement, and lip sync—and how open weights let partners adapt the model.
- fal · Bringing Pickford’s world to life: How story-engine outputs become video prompts, what custom post-training adds, and why dialing back LoRA strength can make characters feel more natural.
- SGLang · One year in, 100 days with H3: Optimizing diffusion inference across data-center and consumer GPUs, and how performance shapes the live experience.
- Pickford · Creative vision and live demo: How audience choices become part of the investigation while the story stays coherent.
🎭 Then, you take control
The session includes a 20-minute live Pickford demo where audience input shapes what happens on screen, followed by a shared Q&A.
Whether you build generative media tools, optimize inference, or create interactive experiences, come see how the pieces fit together
📅 Friday, November 13
🕐 1:50–2:50 p.m. PT
About ****MiniMax
MiniMax is a global AI foundation model company founded in 2022, with the mission “Intelligence with Everyone.” We develop multimodal models spanning text, video, speech, and music, including MiniMax M3.1 Flash and MiniMax H3. Our models power AI products and a developer API platform, helping people build, create, and work in new ways.
Our models and products have cumulatively served over 300 million individual users across 230+ countries and regions, alongside more than 2 million enterprises and developers across 100+ countries and regions.
San Francisco, CA
Venue listed by the organizer · San Francisco, CA. Confirm the final address before traveling.
Open in Google Maps ↗Details come from the event listing and the organizer page; registration is managed by the organizer.
Last checked: Nov 13, 2026. Confirm current availability before making plans.
Organizer page checked: Nov 13, 2026. Check registration for the latest details.