Submissions Now Open for the Story Unboxed – Dreamina Seedance 2.5 3D Reference Creative Contest How many stories can one 3D reference animation generate? This contest invites creators worldwide to submit AI works based on 3D reference. We provide official 3D reference animation assets — and with Dreamina Seedance 2.5 3D white model reference capability, creators can reinterpret the same 3D reference animation to create different characters, styles, worlds, and narratives, letting the same reference blossom into completely different stories. 🏆 Prizes Grand Prizes (jury‑selected, 10 winners): -Champion: $50,000 USD + 1,000,000 Dreamina Credits -Runner‑up: $20,000 USD + 100,000 Dreamina Credits -3rd Place: $10,000 USD + 10,000 Dreamina Credits -4th–10th Place: 1 month of Dreamina Advanced Membership each Social Media Likes Award (Top 10 per platform – TikTok, Instagram, X): Each winner receives 30,000 Dreamina Credits. 📅 Submission Period August 5 – August 31, 2026 📋 How to Participate -Download the official white model reference video, select Dreamina Seedance 2.5, upload the reference, enter your creative prompt, generate your stylized short film, and submit via the Dreamina AI website. -Publish your work on at least one of the following platforms: TikTok, Instagram, or X, with the required hashtags: #DreaminaAI# #DreaminaSeedance25# #DreaminaSeedance25Contest#. -Complete the submission form On Dreamina Website. All three steps must be completed for a valid entry. Tell your story with the same white model. Submit now and show the world your unique vision. #DreaminaAI# #DreaminaSeedance25# #DreaminaAI3DContest#

Today we release LFM2.5-2.6B, an agentic model that runs entirely on-device. It plans, calls tools, and works through multi-step tasks on phones, laptops, PCs, and robots. Data never leaves the device, and the marginal cost of each run is essentially zero. > Pre-trained on ~34T tokens > LFM2.5 flagship hybrid architecture > Context length: 128K > Vocab size: 128K > balanced intelligence per watt > customizable on a single GPU for any specialized task > LFM2 open-weight license Comparable or better scores compared to models up to nearly 4x its size: > ToolSandbox 77.83, ahead of Qwen3.5-9B at 76.44 > Multi-IF 80.07, ahead of Gemma-4-E4B-it at 77.35 > IFStruct 85.49, ahead of Qwen3.5-9B at 78.50 🧵