Updated
Updated
Showing low-relevance items too. Hide low-relevance items
Stability AI@StabilityAIAI score25AIStability AI's interactive research team presented new work at the 19th European Conference on Computer Vision (ECCV) aimed at keeping colors consistent across shots and AI-generated reference photos. The post frames color consistency as a persistent friction point in production.
MiniMax Design (H3)@Hailuo_AIAI score16AIHailuo AI, owned by MiniMax, promotes its MiniMax H3 model as a way to create an army of characters without casting or catering costs. A creator, @koldo2k, says they used GPT-6 Astra-generated work to build a set and story, then animated it with MiniMax H3.
MiniMax Design (H3)@Hailuo_AIAI score12AIHailuo AI shared a sample showing MiniMax H3 can draw with a pencil. A quoted post says a user created an image with MiniMax H3 Design from a prompt, calling the result impressive.
MiniMax Design (H3)@Hailuo_AIAI score9AIThe Hailuo AI Bestiary contest closes tonight, September 14 at 23:59 PT, with $8,000 cash, 200,000 Credits, and ten Audience Choice awards. Entrants must submit a 30-second-or-longer short film generated with MiniMax H3 on any myth, era, or world.
MiniMax (official)@MiniMax_AIAI score36AIMiniMax highlighted open-source community progress on its H3 video generation model, which it built with native stereo audio and multimodal reference control. Recent highlights include FastH3's 4-step distillation running on DGX Spark and Apple Silicon, and NVIDIA's Sol-H3 generating 15 seconds of 768p video with audio in 6.6 seconds on 8×B300 in a warm-inference benchmark. Other releases include VDN's faster-inference attention work with code and weights, and 8-step Acc-LoRAs from Alibaba PAI, with LightX2V offering 4- and 8-step Turbo LoRAs.
Sherwin Wu@sherwinwuAI score40AIGPT Image 2.5 is now available in ChatGPT, and the post says it keeps consistency well while editing images. It also claims state-of-the-art results on all image leaderboards, and suggests users who saw faces shift in GPT Image 2 edits try again.
AIQwen has open-sourced Qwen-Image-2.1, a unified text-to-image generation and image editing model with 7B parameters in its visual generation component. The model can generate regular or transparent RGBA images, supports up to 10 reference images for editing, and is licensed under the Qwen Research License Agreement.
Why it matters: The source specifies the 7B visual component, transparent RGBA output, and up to 10 reference images, which helps readers judge its fit for generation and editing workflows.
Ian Johnson 🔬🤖@enjalotAI score23AIIan Johnson (@enjalot) built a font generator after finding a cluster for each letter of the alphabet in his dataset, with Astra helping write the code. The tool is available as a Hugging Face Space and on GitHub, and the dataset includes SigLIP2 embeddings that allow concept search and clicking a result to jump to similar blocks.
Replicate@replicateAI score46AIOpenAI's GPT Image 2.5 is now available on Replicate, offering precise image editing, sharper details, and higher consistency across multi-turn edits. Two model pages are provided, openai/gpt-image-2.5-sunburst and openai/gpt-image-2.5-flare.
Mark Zuckerberg@finkdAI score5AIMark Zuckerberg announced that the Muse app is available now, and users can start making their own Muse at muse.ai. The post gives no details on what a Muse is or what the app does beyond creation.
World Labs@theworldlabsAI score32AIWorld Labs says its Atlas model maintains 3D consistency during generation by building a spatial context from inputs and previously generated frames. The company adds that real-world images can be added to this spatial context to explore reconstructions of real-world scenes.
Fei-Fei Li@drfeifeiAI score72AIWorld Labs co-founders discuss Atlas, a world model for spatial intelligence, as the key to unifying pixel-level generation and reconstruction. The source says Atlas can digitally capture a 3D representation of a space from three photos, where previously 100 to 300 were needed.
Microsoft AI@MicrosoftAIAI score36AIMicrosoft AI's MAI-Image-2.6 and the new MAI-Image-2.6-Flash are reported to set a new quality/price Pareto frontier for image editing on Artificial Analysis. Both models are now available in Microsoft Foundry.
Mustafa Suleyman@mustafasuleymanAI score25AIMicrosoft's MAI-Image-2.6-Flash is now available in Microsoft Foundry and the MAI Playground. The post links to a Foundry model card for details, but gives no further specifications, pricing, or benchmarks.
Mustafa Suleyman@mustafasuleymanAI score21AIMicrosoft's MAI-Image-2.6-Flash generates images 2x faster than GPT-Image-2, which the post calls the best model in the world. It also uses 72% less GPU, enabling a lower price and what the post describes as the best price-performance score.
Microsoft AI@MicrosoftAIAI score14AIMicrosoft AI says its MAI-Image-2.6-Flash offers the best price-to-quality ratio in the world. The post gives no pricing, benchmark, or speed figures to support the claim.
Google AI Studio@GoogleAIStudioAI score44AIGoogle has made Lyria 3.5, its best-sounding music generation model, available in AI Studio, through the Gemini API, and in the Gemini app. The model produces more expressive vocals and richer musical arrangements, enabling higher-fidelity tracks.
AIMidjourney's alpha site now runs the new v8.2 edit model, with an editor built into the lightbox. Users can edit images with plain-text instructions, attach up to 4 reference images, and view all session edits in one place. The update also adds an early Change Style feature, and the team says speed and error messaging have improved, while drag and drop and the prompt bar are still in progress.
RadixArk@radixarkAI score16AIRadixArk says its latest blog shows how Miles supports multimodal learning through a shared post-training design for vision-language models and diffusion models. The post frames cross-modal learning as necessary for AI to understand and recreate the real world.
Awni Hannun@awnihannunAI score51AIMirai is releasing speculative decoding in its Uzu inference engine, starting with Qwen3.6-27B. The quoted post reports 105 output tokens per second on an Apple M5 Max with 128 GB of unified memory, 2.9× faster than the fastest MLX speculative-decoding implementation Mirai benchmarked. The stack combines DFlash with Mirai's Weaver model, tree-based speculative decoding, Mirai quantization, and Metal kernels for Apple silicon.
World Labs@theworldlabsAI score35AIWorld Labs says its Atlas model places multiple input views within a spatial context to generate image and video frames with precise control. The company demonstrates a hand-designed camera trajectory that produces a one-minute video at 1440p resolution from seven reference images.
World Labs@theworldlabsAI score46AIWorld Labs has introduced Atlas, which it describes as the first multimodal world model that generates image and video frames with pixel-perfect camera control. The model can also reconstruct the generated frames in 3D, letting users move the camera and simulate space and time.
Bryan Catanzaro@ctnzrAI score16AIPeter Kavic from VC Novato explains how DLSS 5 Neural Rendering makes NBA 2K27 look and feel more like a real basketball game. The post shares a YouTube video as the source of this explanation.
Bryan Catanzaro@ctnzrAI score13AIBryan Catanzaro, NVIDIA's VP of applied deep learning research, called DLSS 5 Neural Rendering a major step toward the future of real-time graphics and the culmination of ten years of research. He described the technology as controllable, consistent, and fast, and said he looks forward to seeing what game developers build with it.
AIShanghai AI Lab's InternLM has published Intern Lumina U2, a 16B-parameter MoE model with 1B active parameters that handles text QA, image generation and editing, and image, video, and 3D understanding. The model uses an 8-codebook fully-discrete visual representation built on AToken. Checkpoints are provided for Huawei Ascend NPUs and NVIDIA GPUs under Apache 2.0, with the technical report still listed as coming soon.
Why it matters: The model unifies text, image, video, and 3D understanding with image generation in one framework, a broader scope than single-modality releases.
Midjourney@midjourneyAI score23AIMidjourney has updated its V8.2 edit model to improve image quality. Users who hit problems in the past 24 hours are invited to retry and keep sending feedback, with more updates promised soon.
RadixArk@radixarkAI score12AIRadixArk, a publisher on X, shares a full guide on LoRA supervised fine-tuning for the H3 model in its miles_diffusion repository on GitHub. The post itself contains only a link to the documentation, so no further details about the method or results are given.
RadixArk@radixarkAI score34AIRadixArk introduced LoRA SFT in Miles-diffusion for fast, targeted post-training of diffusion models. The company trained a rank-64 LoRA adapter for MiniMax H3 to improve physical realism, using 254 curated training windows and under 3 hours on 8 GPUs. The adapter can be exported to safetensors and served directly with SGLang without retraining the full model.
Bryan Catanzaro@ctnzrAI score46AINVIDIA's DLSS 4.5 Ray Reconstruction is now available, using a second-generation joint denoiser and super-resolution model. According to the post, it delivers much better image quality at the same compute cost, pushing the trade-off between image quality and rendering cost further.
Replicate@replicateAI score10AIReplicate announces Recraft V4 Styles, which generates and applies style references to produce production-ready raster or SVG vector files. The post says the offer includes 50% off for the first 48 hours.
Replicate@replicateAI score16AIReplicate is letting users try four Recraft V4 variants: styles, styles SVG, styles pro, and styles pro SVG. The post provides links to each model on Replicate, with no details on pricing, benchmarks, or capabilities.
Microsoft AI@MicrosoftAIAI score34AIMicrosoft AI introduced MAI-Image-2.6, its latest image model, which can modify color, style, and visual details while maintaining consistency across iterations. The post positions it as a tool for bringing mood boards to life.
AIGoogle has published TIPS B/14 (v1) on Hugging Face, a contrastive vision-language model that produces spatially rich image features aligned with text embeddings. The model has 86M vision parameters and 110M text parameters at native 448 resolution, and is licensed under Apache 2.0. The release includes usage code for image and text encoding, zero-shot classification, and spatial feature visualization.
Google AI Developers@googleaidevsAI score34AIGoogle shows Gemini 3.7 Flash, paired with Nano Banana and Omni, generating a fully animated parallax landing page from a single prompt. The model calls tools to write the copy, create image references, and render parallax-ready Omni videos.
Microsoft AI@MicrosoftAIAI score32AIMicrosoft's MAI-Image model now ranks third on Arena's image editing leaderboard. The latest MAI-Image model is available in Private Preview on Microsoft Foundry.
Stability AI@StabilityAIAI score38AIStability AI has released two beta tools for Stable Audio 3.0: a plugin that brings audio generation into digital audio workstations (DAWs) and an upgraded experience at StableAudio.com with more editing options. Both are powered by commercially-safe models, so users own their outputs and can distribute them freely. Some features are experimental, and the company says it will keep iterating in real time.
AIStability AI released a Stable Audio plugin that runs Stable Audio 3.0 generation inside DAWs as an instrument, available as a macOS AU and VST3 with Apple Silicon and Intel support. The enhanced StableAudio.com web app adds iterative prompting, audio-to-audio variations, per-track mixing controls, and export, and both tools are in beta and powered by commercially-safe models that users can distribute freely.
Google AI Developers@googleaidevsAI score22AIGoogle AI Developers showcased a Chrome extension built in Antigravity with Gemini 3.7 Flash that generates rich visual definitions when users highlight text while browsing. The post says the tool uses Nano Banana and Gemini Omni to turn the web into an interactive visual encyclopedia, with a demo video linked.
Google Labs@GoogleLabsAI score13AIGoogle Labs is promoting two experimental AI tools: Pomelli, for building marketing campaigns tailored to a brand at scale, and Flow, for crafting visual stories. The post invites users who have not yet tried them to explore both.
AIByteDance has released Bernini-Diffusers-v2 on Hugging Face, a video generation and editing pipeline combining a Qwen2.5-VL planner with Wan2.2 diffusion components. The model card recommends it over Bernini-R for complex requests needing stronger instruction following and multi-step semantic planning. Code and weights are available under Apache License 2.0.