🚀 Awesome news from OpenAI: The GPT-4.1 model family is out! 🚀

OpenAI has announced a new generation of AI: GPT-4.1, GPT-4.1 mini

What has changed compared to GPT-4o?

• ✅ Programming on a new level: GPT-4.1 significantly outperforms previous versions in coding tasks.
• ✅ Better understanding: The models follow complex instructions more accurately and maintain context in long dialogues.
• ✅ Massive context: models support up to 1 MILLION tokens! This allows analyzing huge volumes of code or documents in full.
• ✅ Up-to-date knowledge: The knowledge base is updated through June 2024.
• ✅ Efficiency and cost: The new models run faster and cost less! For example, GPT-4.1 mini matches GPT-4o’s intelligence level but is almost twice as fast and 83% (!) cheaper.

What is the difference between the new models?

• GPT-4.1: The flagship for the most complex tasks (coding, research).
• GPT-4.1 mini: The optimal balance of power, speed, and cost. Supports text and images.

Now the result size can be up to 4096*4096 pixels, provided that the mask size does not exceed 30% of the image size.

Now, in the task history, you can see the current progress of processing long video files.

This error occurred during batch processing of images (more than 8 images).

New interface:

• You can change a face in any photo.
• Face selection feature (serial number).
• Improved face quality and increased output resolution.

Google has released Gemini 2.5 Pro — a powerful AI that handles tasks related to logic, mathematics, and coding. Its capabilities were assessed using challenging tests, such as GPQA — a set of graduate-level questions in physics, chemistry, and biology, and AIME — a mathematics competition for high school students. These tests demonstrate how well the AI solves problems that are difficult even for experts.

We have updated our AI background removal model. It now works even more accurately.

LuminaGen is the latest development in AI-based text-to-image generation. The model leads in lighting and overall aesthetics, surpassing competitors such as DALL-E 3 and Midjourney. It also demonstrates impressive versatility across various artistic styles, from photorealism to impressionism, and from abstract art to anime-style illustrations.

Note: censorship is present.

PhotoStab v4. Available in two versions:

• PhotoStab Turbo - when you need it fast, inexpensive, without requiring maximum quality or for experiments.
• PhotoStab - image drawing with maximum quality.

What's new:

• Ability to draw text.
• No censorship.
• The original image is no longer redrawn.

• The new model from Perplexity - Sonar Deep Research can automatically create comprehensive analytical articles by processing a huge amount of source data from the internet.
• Support for Epub files
• Price reduction on all models, especially on Google Gemini Pro

PhotoStab v4. Available in two versions:

• PhotoStab Turbo - when you need it fast, inexpensive, without maximum quality or for experiments
• PhotoStab - editing with maximum quality.

What's new:

• Ability to draw text. For example, highlight a t-shirt and write in the prompt: text "Bad Cat" on a t-shirt.
• No censorship.
• The original image (outside the mask) is no longer redrawn.

What's new in Manga Master v2:

• Our AI-based anime image generator now supports natural language processing (NLP), eliminating the need for complex tagging systems — just describe what you want to create!
• Add custom text elements to your anime art. Specify the appearance of the text with natural descriptions. Just describe how you want the text to be visualized, for example, yellow font in a pastel pink speech bubble.

Meet Waifu Studio - high-quality, detailed images in anime style. The model responds well to complex prompts (text descriptions of the desired image), allowing users to achieve more accurate results.

The Velvet Render model creates highly realistic images with photographic quality, excelling particularly in rendering human portraits and detailed environments.

Key features:

• Photorealistic results: the model produces strikingly lifelike images, with a particular strength in human portraits and detailed surroundings.
• Minimal prompts: users report impressive results with simple, brief requests, making the model accessible for both novice and experienced users.
• Diverse subjects: while the model specializes in depicting people, it can generate a wide range of content, including landscapes, animals, and objects with high accuracy.
• Customizable aesthetics: the model responds well to stylistic modifiers, allowing users to adjust the mood, lighting, and overall aesthetics of the generated images.

1. Improved quality and realism of images, especially when generating photorealistic and detailed 3D scenes.
2. Increased accuracy in aligning with text prompts, allowing for more relevant results.
3. Enhanced styling capabilities, including a variety of artistic styles and effects.
4. Improved handling of complex scenes with multiple objects.
5. More representative generation of images of people with various skin tones and facial features without the need for detailed prompts.
6. Enhanced visualization of human anatomy compared to General XL.

• Emotional intelligence: More "human-like" communication.
• Improved accuracy: The model better understands requests and provides more accurate answers.
• In-depth understanding: More knowledge leads to better solutions.

1. The bug that caused the dialogue context to be lost when using the OpenAI model has been fixed.
2. Anthropic has announced Claude 3.7 Sonnet, its most advanced language model, featuring a hybrid reasoning method. It offers both rapid responses and the capability for in-depth analysis ("thinking"). Claude 3.7 Sonnet demonstrates significant improvements in coding and web development. Already available in PrboChat.
3. Links to information sources are now displayed when using the Perplexity model.

A cutting-edge model for creating videos that transforms text and images into high-quality clips, characterized by realism and smooth motion. It provides improved scene understanding, natural lighting, and cinematic camera movements, making it suitable for professional marketing and social media content creation.

• The new version of the PicRefiner algorithm enhances image quality by reducing noise and improving contrast, color range, and sharpness.

• New internet access models from Perplexity based on DeepSeek
• The latest development from xAI - the Grok beta model

OpenAI has introduced O3 Mini, its new language model that delivers faster and more accurate responses compared to its predecessor, O1 Mini, reducing response time by 24% to 7.7 seconds. It is optimally utilized for various tasks in the fields of science, technology, engineering, and mathematics (STEM).

Fixed bugs:

• Video clips were not playing on some devices and the rewind function was not working.
• Incorrect link to the video file in email notifications.

An low-cost model for solving complex problems through thinking

Increased from 512 to 1024 pixels.

What's new:

• Improved detail.
• Universal. From photorealism to anime.
• Added setting - creativity.

Main characteristics:

• Resolution: approximately 1300 x 700
• Frames per second: 24
• Duration: 5 seconds.
• Cost: $0.7
• Fast generation.

Thanks to the new large model, the descriptions are more detailed and in a natural style.

Enter a description of the video in any language and receive a unique 6-second video clip. You can also enhance the video to Full HD 50 FPS from the context menu of the result.

Key features:

• Support for more styles.
• Excellent at drawing people, including hands.
• Easily handles text.
• Fast generation.

Main improvements:

• Hair looks softer, with more distinct individual strands
• Skin texture has become more realistic
• Easier to create images of full-length people in various poses
• The image quality of people and objects in the background has noticeably improved

Capabilities:

• increase resolution to FullHD
• reduce noise, add details
• enhance faces
• increase video frame rate to 60 frames per second

Can extend (augment) your photo. You just need to choose directions and write a prompt.

https://problembo.com/services/image-extender

Removes both small and large objects by completely redrawing a part of the image (fills in the removed areas).

Changes in app functionality:

• The latest algorithm for high-quality background separation even in challenging cases.
• Processing up to 20 files at a time.

The PicRefiner algorithm has been added. A distinctive feature is the automatic adjustment of quality enhancement settings for a specific image.