10 AI Tools That Stole the Show This Week!

The AI providers being launched or introduced proceed to shock us each week. We’ve compiled essentially the most notable ones of the week for you.

HUNYUAN3D 3.0

Tencent launched Hunyuan3D 3.0, providing 3 occasions increased precision and 1536° extremely HD voxel modeling. It will probably seize lacking particulars, practical facial options, and professional-grade textures for gaming, movie, and e-commerce functions. https://hunyuan-3d.com/

WAN2.2

Wan launched Wan2.2, a 5B parameter video diffusion mannequin with an MoE structure that gives increased capability for a similar price. It will probably ship cinema-quality visuals, complicated movement technology, and environment friendly 720p text-to-video and image-to-video output at 24 fps. https://wan.video/

MOONDREAM 3

Moondream 3 launched as a 9B parameter, 2B lively MoE visual-language mannequin, offering state-of-the-art visible reasoning in a compact, application-friendly design. https://moondream.ai/

SRPO

Tencent-Hunyuan unveiled SRPO, a diffusion fine-tuning methodology that stabilizes the coaching course of, corrects noisy photos, and shortens computation time. This methodology permits quicker optimization, prevents reward hacking, and helps controllable model changes for fashions like FLUX.1.dev. https://www.srpo.web/

REVE IMAGE

Reve launched Reve Picture, which mixes picture technology, restyling, a drag-and-drop editor, a artistic assistant, and a beta API. Customers can create and edit photos with pure language and combine Reve’s capabilities into their very own functions. https://app.reve.com/

LING-FLASH 2.0

Ling-flash-2.0 is now open-source and is a 100B parameter MoE LLM with 6.1 billion lively parameters. Skilled on 20T+ tokens, it displays near-perfect efficiency in complicated reasoning, code technology, and frontend improvement, making it essentially the most superior amongst dense fashions underneath 40 billion parameters. https://huggingface.co/inclusionAI/Ling-flash-2.0

VOXCPM

VoxCPM, a tokenizer-free TTS mannequin powered by MiniCPM-4, presents zero-shot voice cloning and hyper-realistic speech with pure concord. Skilled with over 1.8 million hours of information, it achieves state-of-the-art efficiency. https://voxcpm.com/

UMO

UMO, a unified multi-identity optimization framework for picture customization, was launched. It will probably guarantee excessive id consistency, cut back entanglement amongst a number of reference photos, and will likely be absolutely open-source with fashions, scripts, and coaching code. https://bytedance.github.io/UMO/

RAY3

Luma AI launched Ray3, the primary reasoning video mannequin able to producing studio-quality HDR. Its new draft mode permits quick iteration with improved physics and coherence and is now accessible free of charge in Dream Machine. https://lumalabs.ai/ray

PAPER2AGENT

The newly introduced Paper2Agent infrastructure routinely converts educational papers into lively AI brokers. Utilizing a number of sub-agents, the system builds a strong Mannequin Context Protocol (MCP) from a paper’s textual content and code, enabling the ensuing agent to use the paper’s strategies and information to new tasks. https://github.com/jmiao24/Paper2Agent

You May Additionally Like;

Copy URL
URL Copied