←Image generation
← All topics

Image generation

Image generation, editing, visual tools and models.

Public reading groups
53
Last 30 days
25
Directory status
Curated topic

LatestMicrosoft Releases Image Editing Model MAI Image 2.7, Ranked 3rd on Arena Edit

Latest reading groups

Duplicate reports of the same fact are grouped

  1. Microsoft AI BlogFirst-partyModel releases
    AI score67

    Microsoft Releases Image Editing Model MAI Image 2.7, Ranked 3rd on Arena Edit

    Microsoft AI has released the image editing model MAI Image 2.7, launched in preview form, ranking 3rd on Arena Edit and 2nd on the Artificial Analysis image editing leaderboard, making Microsoft AI the 2nd-ranked lab. The company states that the model outperforms 2.6 on all 10 use cases and 7 categories of editing actions evaluated by Artificial Analysis, leading the next-best model by 24 Elo in scenes and styles. It is currently available for trial on Arena, Artificial Analysis, and MAI Playground, with text-to-image functionality coming soon.

    Blurry, light-colored flowers against a blue sky.

    Why read thisThe official rankings of this image editing model on Arena Edit and Artificial Analysis leaderboards are given, along with the improvement margin of use cases relative to 2.6.

  1. Krea AIFirst-partyProduct release
    AI score62

    Krea releases Krea 3 image editing feature

    Krea has released Krea 3, which supports fast editing of any local region of an image, with generated resolution up to 2K. The company says end-to-end 2K editing takes under 8 seconds, and it offers bounding box control and precise multi-turn editing, with image quality benchmarked against Nano Banana Pro or GPT Image 2.5. The feature is now available in beta.

    Why read thisThe official release provides benchmark information on Krea 3's editing speed, resolution, and image quality, making it easier to assess changes in image editing workflows.

  1. Unite.AIModel releases
    AI score79

    Canva Releases Its In-House Image Model Canva Craft

    On October 22, Canva released Canva Craft, the first foundation model trained from scratch by its internal lab Canva Research. It is now available to Canva users worldwide through Dream Lab and to developers via a public API. The model uses a two-stage architecture: first a planning model parses the prompt and generates a composition plan, then a rendering model outputs images at up to 2K resolution supporting more than 30 aspect ratios, and can generate text in 55 languages. Canva says Craft debuted fourth on DesignArena, and previews that two faster variants, Craft-fast and Craft-ultrafast, are coming soon.

    Why read thisThe original article breaks down Craft's two-stage architecture of first planning the composition and then rendering, and how it serves editable layers, making it easier to understand the differences of design-oriented image models.

  1. MediaNama
    AI score76

    TTP investigation: teen accounts with parental controls enabled can still download undressing AI apps from Google and Apple stores

    A Tech Transparency Project investigation using four test accounts for 13-year-old boys found that even with parental controls enabled, 135 AI apps capable of generating nude images could still be accessed on the Apple and Google app stores. Both platforms rely on developers' self-reported questionnaires to determine age ratings, causing apps with sexually explicit features to receive ratings suitable for children; as of September 28, Apple had removed 25 apps and Google had removed 6, but all the removed apps remained fully usable on the test phones.

    Why read thisThe investigation shows that app store age ratings rely on developers' self-reporting, and neither parental controls nor takedown measures have blocked such apps.

  2. Google ResearchFirst-partyResearch
    AI score64

    Google Research uses AI to assist blind-sweep ultrasound assessment of gestational age, with error comparable to experts

    Google Research published a study in JAMA Network Open; its AI gestational age assessment model based on blind-sweep ultrasound achieved a mean absolute error of 4.18 days in prospective testing in Chicago (n=192) and Nairobi (n=193), reaching non-inferiority compared with expert sonographers' 4.48 days. The model was fine-tuned only on data from 120 participants in Chicago and did not use Nairobi data; the errors in the two locations were 4.10 days and 4.27 days, respectively. The study used a Clarius C3 handheld probe, with standardized blind sweeps performed by operators without professional ultrasound training.

    Bar chart comparing the mean absolute error between the AI model and expert clinical standards across multiple cohorts.

    Why read thisThe study prospectively validated the same model in Chicago and Nairobi, maintaining the error level without local fine-tuning, showing how cross-scenario generalization can be implemented.

  1. falModel releases
    AI score68

    fal claims FLUX 3 Max scored 1,139 Elo in third-party evaluations, ranking in the top tier of image-to-video

    fal states that FLUX 3 Max achieved 1,139 Elo in third-party benchmark evaluations, entering the top tier of frontier image-to-video models based on independent human preference assessments. The model was launched through a collaboration between fal Research and bfl_ai; earlier introductions said its image-to-video quality tied for first in independent third-party benchmarks, and it can generate 10-second, 720/768p videos in about 12 seconds. fal Research says it is continuing post-training of the model and will subsequently improve quality and performance.

    Why read thisThe official Elo score and generation speed of FLUX 3 Max in third-party human preference evaluations are provided, which can be used to judge its position among frontier image-to-video models.

  1. AYi AI NotesAnalysis
    AI score83

    3D modeling comparison test between DeepSwarm V4.1 Flash and Claude Opus 5.5

    The author compared the 3D modeling outputs of DeepSwarm V4.1 Flash and Claude Opus 5.5 using the same prompt, and questioned whether the two are at the same level in lighting and floating island details. The tweet includes a comparison video, but no further test conditions or quantitative results are given.

    Why read thisThe author compared the 3D modeling outputs of the two models using the same prompt, so readers can observe the differences between them in lighting and floating island details.