OpenAI Releases GPT-4.1 API

Welcome to another edition of AI PlanetX.

OpenAI launches GPT-4.1 for coders; ByteDance's Seaweed leads in AI video; Google decodes dolphin speech.

• Hottest AI News

• Top AI & SaaS Tools

• Top AI & Tech News

• Interesting Uses of AI: A Prompt to Change Your Life

• AI Video Tutorial

• F-R-E-E AI Course of the Day: Generative AI with LLMs

Hottest AI News

OpenAI

OpenAI Releases GPT-4.1 Models Optimized For Coding Tasks

OpenAI has released GPT-4.1, a family of three models built for coding and instruction. Available through the API, these models aren't yet in ChatGPT, as Google and Anthropic continue developing their own coding-focused AIs.

Details:

• Lineup includes GPT-4.1, mini, and nano, all with a 1-million-token context (750,000 words). GPT-4.1 excels at coding, while mini and nano are faster and cheaper

• OpenAI improved these models based on developer feedback, enhancing coding, format, and consistency. GPT-4.1 scored 52% to 54.6% on the SWE-bench, below Gemini 2.5 Pro (63.8%) and Claude 3.7 Sonnet (62.3%)

• Reliability decreases as input tokens approach 1 million, with accuracy dropping from 84% at 8,000 tokens to 50% at maximum. GPT-4.1 is more literal than GPT-4o and needs clearer prompts

This release advances OpenAI's goal of an "agentic software engineer" for complete app development. GPT-4.1, with a June 2024 knowledge cutoff, offers better current understanding.

Find out why 1M+ professionals read Superhuman AI daily.

In 2 years you will be working for AI

Or an AI will be working for you

Here's how you can future-proof yourself:

1. Join the Superhuman AI newsletter – read by 1M+ people at top companies

2. Master AI tools, tutorials, and news in just 3 minutes a day

3. Become 10X more productive using AI

Join 1,000,000+ pros at companies like Google, Meta, and Amazon that are using AI to get ahead.

Sign up and start learning AI

ByteDance

ByteDance Seaweed Emerges as Superior AI Video Tech

ByteDance just launched Seaweed, a video model with 7B parameters that competes with much larger models, challenging the "bigger is better" approach in generative AI.

Details:

• Seaweed matches or exceeds larger models like Kling 1.6, Google Veo, and Wan 2.1, especially in image-to-video tasks, even outperforming Sora. It supports text-to-video, image-to-video, and audio-driven synthesis with 20-second outputs

• Includes advanced features like multi-shot storytelling, controlled camera movements, and synchronized audio-visual generation, with a focus on realistic human animation and lip syncing

• Chinese tech companies, led by ByteDance's Seaweed, Alibaba's Wan, and Kling, are dominating the video AI field, showcasing China's emerging leadership

Seaweed's impressive performance with minimal compute requirements suggests we've reached an inflection point where efficiency, not just scale, can drive AI innovation.

DeepMind

Google Develops AI That Interprets Dolphin Communication

For decades, scientists have sought to understand dolphins' clicks and whistles. Recently, Google, Georgia Tech, and the Wild Dolphin Project are collaborating on DolphinGemma, an AI model to decode dolphin communication.

Details:

• WDP has studied Atlantic spotted dolphins in the Bahamas for nearly 40 years, linking specific sounds—like signature whistles and squawks—to behaviors such as mother-calf reunions and fighting

• Google's DolphinGemma, using SoundStream technology, identifies patterns in dolphin sounds much like language models for humans. It runs on Pixel phones to detect recurring patterns and explore meaning in dolphin communication

• The CHAT system uses synthetic whistles associated with toy dolphins, teaching dolphins to mimic sounds to request objects. Pixel smartphones help with recognition, aiming to create a basic two-way vocabulary that expands with more decoded dolphin sounds

With AI-powered tools like DolphinGemma and innovative tech in the field, scientists are closer than ever to understanding—and maybe even talking with—dolphins.

Top AI & SaaS Tools

• reap (Life-time Deal): AI tool that lets you turn long-form videos into clips perfectly captioned and formatted for every major social media platform

• SpreadSimple: Turn your Google Sheet into a website in minutes—update the sheet, and your site updates automatically [F-R-E-E]

• Prompt Catalyst: Prompt crafter dedicated to developing precise prompts specifically for video and image generation [F-R-E-E]

• Hera: Transform any text into fully customizable motion graphics [F-R-E-E]

• Andi: AI search engine prioritizing privacy, delivering direct answers with no ads or tracking [F-R-E-E]

Top AI & Tech News

• NVIDIA plans to build up to $500B in AI infrastructure in the U.S. over the next four years

• Hugging Face enters robotics by acquiring Pollen Robotics and its open-source humanoid, Reachy 2

• Apple is developing an AI training method to improve its models without compromising user privacy

• LM Arena launched Search Arena Leaderboard, with Google’s Gemini-2.5-Pro and Perplexity’s Sonar topping the list

• OpenAI plans to release two models this week, o3 and o4-mini, designed to generate scientific ideas and automate high-level research tasks

Interesting Uses of AI

AI Art Spotlight

Model: Midjourney V7

Prompt:

a beautiful young Korean woman gently resting her face against the side of a majestic white horse, both facing forward, intimate and serene moment, soft dreamy lighting, romantic fine art digital painting style, delicate brushstrokes, ethereal atmosphere, cinematic composition, highly detailed --chaos 50 --ar 2:3 --stylize 750 --v 7

AI Prompt of the Day

Top AI Video Tutorial

This is the Most Powerful Way to Use AI

Complimentary AI Course of the Day

Generative AI with Large Language Models

This course covers foundational knowledge, practical skills, and how generative AI works, guided by AWS AI experts, while exploring the latest research and real business applications.