UncensoredHub.ai

Image models Video models LoRA models LLM models

Prompts Gallery Journal ToolsNEW

Newsroom

Staleness-Adaptive Trust Region cuts asynchronous RL performance loss to 3% at 8× policy lagReleases · 7d ago Distilled RL transfers knowledge across model families without unconditional imitationResearch · 7d ago Qwen-Music generates full vocal songs from text and lyricsReleases · 7d ago LongStraw trains RL models at 2.1M tokens on eight H20 GPUsReleases · 7d ago ShortOPD cuts pruned LLM recovery time by 75% while raising generation quality 9×Releases · 7d ago Claude Design launches as Anthropic Labs visual collaboration toolPlatform · 7/14/2026

View all news →

Image models Video models LoRA models LLM models

Prompts Gallery Journal News Polls Free Tools Agents Can I run it?

ZenCreator

Pro-grade AI content creation. Image, video, face-swap, lipsync, and upscaling behind one API.

Loading…

UncensoredHub.ai© 2026

contact@uncensoredhub.ai

Image Video LoRA LLM Prompts Gallery Journal News Polls Tools Agents

About Terms Privacy Cookies DMCA

18+ · Adults only · Not affiliated with hosted platforms

← Back to hub

News · 233 articles

What's happening

Releases, platform updates, funding, industry shifts, community work, and research from across the AI scene.

All Releases Platform Funding Industry Community Research NSFW

Hugging Face adds one-click SageMaker Studio deploy button

ReleasesPlatformNSFW

Hugging Face adds one-click SageMaker Studio deploy button

Hugging Face model cards now include a direct deploy-to-SageMaker Studio button, letting AWS users spin up inference endpoints without leaving the browser.

Alberta taps Claude for cybersecurity threat analysis in public-sector first

IndustryPlatformNSFW

Alberta taps Claude for cybersecurity threat analysis in public-sector first

The Government of Alberta is using Anthropic's Claude to support cybersecurity operations across provincial systems, marking one of the first public-sector deployments of a frontier language model in critical infrastructure defense.

Jul 12, 2026

Half of Fortune 500 now use open-source AI, Hugging Face CEO says

Half of Fortune 500 now use open-source AI, Hugging Face CEO says

Clem Delangue says roughly half the Fortune 500 now use Hugging Face to download and deploy open-source models, marking a decisive turn from closed API dependencies.

LingBot Video 1.3B and LingBot-World-v2 14B land on ComfyUI with 8GB VRAM support

LingBot Video 1.3B and LingBot-World-v2 14B land on ComfyUI with 8GB VRAM support

Two LingBot models—a 1.3B video generator and a 14B world-generation model—have been repackaged as ComfyUI custom nodes with GGUF quantization, enabling local video and scene synthesis on consumer GPUs.

Qwen3-24B abliterated weights hit HuggingFace with 256k context

Qwen3-24B abliterated weights hit HuggingFace with 256k context

Higashikawa released abliterated GGUF weights for Qwen3-24B-A4B-Freedom-Thinking, a mixture-of-experts model with 256k context and no safety filters.

Qwen 3.5 9B uncensored writer fine-tunes land in GGUF quantizations

Qwen 3.5 9B uncensored writer fine-tunes land in GGUF quantizations

Two new GGUF-quantized variants of Qwen 3.5 9B fine-tuned for creative writing without safety filters landed on HuggingFace this week, both targeting local inference via llama.cpp.

Seedance 2.5 video model costs 50% more per second than 2.0

IndustryNSFWPlatform

Seedance 2.5 video model costs 50% more per second than 2.0

Unverified reports suggest ByteDance's Seedance 2.5 video model costs roughly 1.5× more per second than version 2.0, with 720p generation priced around $5 for 30 seconds through subscription credits.

Midjourney demands studios disclose their own AI use in copyright lawsuit

IndustryNSFWPlatform

Midjourney demands studios disclose their own AI use in copyright lawsuit

The image generator is pushing three studios to reveal their own AI practices as part of an ongoing lawsuit, escalating transparency demands in the entertainment industry's copyright battles.

OpenAI publishes defense partnership framework with democratic oversight pledges

IndustryNSFWPlatform

OpenAI publishes defense partnership framework with democratic oversight pledges

OpenAI released a framework for working with government and defense agencies, emphasizing democratic accountability, responsible AI use, and public safety guardrails.

Anima-Bakes fine-tune launches on HuggingFace as unrestricted Circlestone derivative

ReleasesNSFWPlatform

Anima-Bakes fine-tune launches on HuggingFace as unrestricted Circlestone derivative

KtiseosNyx released Anima-Bakes, an open-weight fine-tune of Circlestone Labs' Anima base model, flagged as not-for-all-audiences on HuggingFace.

Half of Fortune 500 companies now use Hugging Face for open-source AI

Half of Fortune 500 companies now use Hugging Face for open-source AI

Clem Delangue reports that roughly 250 Fortune 500 companies now download models and datasets from Hugging Face, as enterprises shift from closed APIs to open weights for cost control and customization.

RQSA-MIL v1 ONNX classifier for content moderation debuts on HuggingFace

RQSA-MIL v1 ONNX classifier for content moderation debuts on HuggingFace

Shivapreetham17 released an ONNX image-classification model for content moderation on HuggingFace, marked not-for-all-audiences with non-standard licensing.

Meta Superintelligence Lab releases Muse Image and Video with code-generation and self-correction

ReleasesNSFWPlatform

Meta Superintelligence Lab releases Muse Image and Video with code-generation and self-correction

Meta's renamed research division released Muse Image and Muse Video, successors to Emu that integrate diffusion and LLM layers to write Python code, parse web references, and self-correct generation errors.

Proactive memory agent lifts long-horizon task pass rates by 8.3 pp on Terminal-Bench

ReleasesResearchNSFW

Proactive memory agent lifts long-horizon task pass rates by 8.3 pp on Terminal-Bench

A new memory module runs alongside unmodified action agents, deciding when to inject context reminders and when to stay silent, beating passive retrieval across two benchmarks.

Qwen3.5-27B abliterated fine-tune ships on HuggingFace for unrestricted multimodal inference

Qwen3.5-27B abliterated fine-tune ships on HuggingFace for unrestricted multimodal inference

DavidAU released abliterated weights for Qwen3.5-27B-Polar-Rev1, a 27-billion-parameter multimodal model stripped of safety tuning and available for unrestricted image-text-to-text generation.

LTX Director bridges Blender staging with diffusion-based video rendering

LTX Director bridges Blender staging with diffusion-based video rendering

A new tool lets filmmakers stage scenes in Blender and render them through generative video models, using depth maps and motion vectors instead of text prompts to control camera, lighting, and character movement.

OpenAI's No. 2 executive Fidji Simo steps down as AGI chief

OpenAI's No. 2 executive Fidji Simo steps down as AGI chief

Simo transitions to part-time advisor role after neuroimmune condition extends her medical leave, leaving OpenAI's second-ranking position vacant ahead of a possible IPO.

Plan A: Former OpenAI researchers propose international AI transparency treaty to manage superintelligence race

Plan A: Former OpenAI researchers propose international AI transparency treaty to manage superintelligence race

Former OpenAI researchers released Plan A, a policy scenario mapping a coordinated path to superintelligence through mandatory transparency in frontier AI research and an international agreement to replace the current arms race.

1 2 3 4 5 6 7 8 9 10 11 12 13 Next →