Skip to content
Drix10 Blog

DeepSeek v4.1 Flash 1017 is top open-source model

, 5 items in AI Holodeck and Virtual Worlds, 3 min read

In this digest (5 items)

DeepSeek v4.1 Flash 1017 is presented as the top open-source model. It claims Opus 5 level intelligence, complete privacy, higher speed, and 95% lower cost.

Key points

  • Model: DeepSeek v4.1 Flash 1017

  • Claim: Opus 5 level intelligence, completely private, way faster, and 95% cheaper

Sources

Tesla Roadster Sandbox live with 5 maps and demos

Tesla Roadster Sandbox is live now with 5 maps and demos. Built on usecrayon with opus 5.5 and Threejs.

Key points

  • Live: 5 maps and demos available

  • Platform: usecrayon with opus 5.5 and Threejs

Sources

Sonilo – First video soundtrack tutorial in 2-minute walkthrough

Sonilo offers a 2-minute walkthrough to create music and sound effects for a video. The tutorial shows uploading a video, generating a brief, creating music and sound effects, assigning music per scene, and exporting the result.

Key points

  • 2-minute walkthrough demonstrates adding music to a video with Sonilo.

  • Steps include upload, brief generation, music and sound effects creation, scene-wise music setting, then export.

Sources

Sonilo – added as a ChatGPT plugin for music creation

Sonilo is now available as a ChatGPT plugin. Install it from Plugins, select it in your chat, and describe what you need, from lo-fi beats to cinematic whooshes. It creates music and sound effects without leaving the chat.

Key points

  • Install Sonilo from Plugins and select it in your chat

  • Describe what you need, from lo-fi beats to cinematic whooshes, to create music and sound effects directly in ChatGPT

Sources

OmniHuman 2.0 demo shows single image to talking avatar

OmniHuman 2.0 turns a single portrait into a lifelike, expressive talking avatar. The demo includes audio-driven generation, natural gestures, emotional expressions, multi-person performance, precise camera and motion control, native 720p output, and up to 5 minutes of video.

Key points

  • Feature: audio and text-driven generation of dynamic avatar videos.

  • Capability: native 720p output and up to 5 minutes of video length.

Sources