|
In this newsletter:
- Anthropic launches Claude Sonnet 5 with Opus-level agents at lower cost
- Google's Nano Banana 2 Lite generates images in 4 seconds at minimal cost
- Runway adds Gemini Omni Flash for chat-based video editing
Plus, you’ll find new AI tools and this week’s top AI news headlines!
🤖 Anthropic Launches Claude Sonnet 5 With Opus-Level Agents At Lower Cost
Anthropic released Claude Sonnet 5 as the new default model for Free and Pro plans, closing the gap with Opus 4.8 on agentic tasks while keeping costs lower.
What's new:
- Available now across all plans as the default for Free and Pro, available to Max, Team, and Enterprise users
- Introductory pricing of $2 per million input tokens and $10 per million output tokens through August 31, 2026, then moves to $3/$15 per million tokens
- Substantial improvement over Sonnet 4.6 on reasoning, tool use, coding, and knowledge work
- Performance close to Opus 4.8 on agentic tasks like browsing, computer use, and complex multi-step work
- Finishes complex tasks where previous Sonnet models would stop short
- Checks its own output without being explicitly asked
- Lower hallucination and sycophancy rates than Sonnet 4.6
- Better at refusing malicious requests and resisting prompt injection attacks
- Available in Claude Code and Claude Platform
- Rate limits increased across Chat, Cowork, Claude Code, and Claude Platform
Tokenizer note:
- Uses an updated tokenizer that can increase token count by 1.0 to 1.35x, depending on content
- Introductory pricing set to make transition roughly cost-neutral compared to Sonnet 4.6
Cybersecurity safeguards:
- Cyber safeguards are enabled by default, the same as Opus 4.7 and 4.8
- Could not develop working exploits in Firefox vulnerability tests, slightly higher partial success rate than Sonnet 4.6
- Substantially weaker on dangerous cyber tasks than Opus 4.8 and Mythos 5
- Part of the Cyber Verification Program for organizations needing reduced guardrails
Early partner feedback:
- Testers described it finishing complex tasks where previous Sonnet models stopped short
- Checks its own output without being asked
- Described as doing agentic work at an attractive price point
Why it matters:
Sonnet 5 shifts the value equation for anyone paying Opus prices to get reliable agentic performance - the same kind of multi-step, self-checking work that used to require the most expensive model is now available at roughly half the cost during the introductory period.
For businesses running high-volume AI workflows, that pricing difference compounds quickly, and Anthropic is clearly signaling that Sonnet is now the practical choice for most production use cases rather than a budget compromise.
👀 Read more about Claude Sonnet 5!
🎨 Google's Nano Banana 2 Lite Generates Images in 4 Seconds at Minimal Cost
Google released Nano Banana 2 Lite, a faster and cheaper version of its AI image generator, built for high-volume workflows that require images to be produced quickly and at scale.
What's new:
- Available now through Google AI Studio, Gemini API, and Gemini Enterprise Agent Platform
- Generates images in four seconds, optimized for rapid iteration and large volume production
- Costs $0.034 per 1,000 images, making it significantly cheaper than existing options
- Replaces original Nano Banana, which Google now refers to as its legacy model
- Positioned as a high-speed option, while Nano Banana 2 remains the generalist choice, and Nano Banana Pro handles advanced use cases
Also announced the same day:
- Gemini Omni Flash's wider release at $0.10 per second of video output
- New demo app Omni Product Studio takes static images and transforms them into cinematic e-commerce videos
- Google described the two models together as tools for building "end-to-end multimedia experiences that connect rapid image generation with video creation and editing"
Why it matters:
The four-second generation time and near-zero cost per image shift AI image generation from a creative tool into a production pipeline ingredient, making it practical for businesses that need hundreds of product images, ad variations, or content drafts without the cost adding up.
Pairing it with Gemini Omni Flash on the same day signals Google is building a full content production stack where images flow straight into video, which is exactly the workflow e-commerce and marketing teams currently stitch together manually across multiple tools.
👀 Read more about Google Nano Banana 2 Lite!
🎬 ByteDance Launches AI Video Model That Creates Longer, Cheaper 4K Videos
Runway added Google DeepMind's Gemini Omni Flash to its platform, allowing paid users to generate and edit videos from prompts, images, or existing footage via natural-language conversation.
What's new:
- Available now for paid Runway users
- Generate a video starting from a text prompt, image, or existing video clip
- Edit finished clips through conversation without regenerating from scratch
- Model maintains character, scene, and lighting consistency across multiple edits
- Independent tests show consistency holds for up to four consecutive edits before minor drift
- Native audio generated in the same pass as video, no separate audio step needed
- Clips run 3 to 10 seconds at 720p in landscape or portrait format
- Priced at $0.10 per second of generated video
- Every clip includes a SynthID watermark and C2PA metadata identifying AI-generated content
- Model won't lip-sync a real person's photo to audio, limiting deepfake potential
How it compares:
- Behind Veo 3.1 on raw output quality and maximum resolution (no 4K option)
- Stronger than competitors on conversational editing, most other models break consistency on the second edit
- Runway's own Gen-4.5 and Aleph 2.0 remain available for users who need higher output fidelity or longer clips
Why it matters:
Most AI video tools still work the same way: generate a clip, hate something about it, start over. Gemini Omni Flash changes that by letting you send a note instead of booking a reshoot - say "change the lighting" or "swap the jacket," and only those frames get re-rendered while everything else stays intact.
For content teams making product videos or social content at volume, that's the difference between a tool you actually build a workflow around and one that's too slow to use consistently.
👀 Read more about Runway's Gemini Omni Flash!
My Latest LinkedIn & X/Twitter Posts:
- ChatGPT vs Gemini vs Claude vs Grok vs DeepSeek (view post)
- 9 ChatGPT prompt frameworks for expert-level results (view post)
- 5 proven ChatGPT prompt frameworks for better results (view post)
- Guide for matching your task with the best AI tool (view post)
- 15 ChatGPT prompts to structure your research from idea to output (view post)
In partnership with Kit:
As a creator, your time should be spent doing what you love, not juggling a dozen tools just to run your business.
Kit gives you everything you need in one place:
✅ Build and grow your email list (I use Kit for my newsletter)
✅ Easily monetize with paid newsletters and digital products
✅ Automate your emails with triggers and custom workflows
✅ Track what’s working and optimize with powerful insights
It’s not just email. It’s your entire creator business, simplified.
Join a thriving community of successful creators.
Use Kit for Free — and Start Building Smarter
New AI Tools to Boost Your Productivity:
- Vivix: Creates videos from text prompts.
- Marvlus: Audits websites for SEO and growth opportunities.
- Amaiko: Automates daily work with an AI assistant.
- LDOO: Answers marketing questions from your data.
- DoJob: Finds jobs and automates your job search.
- Gravytrainer: Builds AI chatbots from your content.
- Website Auditor: Audits AI visibility, SEO, and website performance.
- FastBots: Creates AI chatbots for websites.
- RiffKit: Recreates TikTok videos with your own content.
- Intuo: Connects you with experts for one-on-one calls.
- Vidu: Turns images into videos.
- Stigg: Controls AI usage, credits, and access.
- areyoufoundbyai: Tracks your brand's visibility in AI search.
- Sacha: Creates and schedules LinkedIn and X posts.
- RankBits: Tracks your brand's visibility in AI search.
- ApplyBtn: Autofills job applications.
- Cewsco: Automates business tasks with AI.
- Velma: Analyzes voice conversations for emotion and behavior.
- Audio to Text Converter: Transcribes audio and video into text.
This Week's Top AI News Headlines:
- Google NotebookLM Can Now Turn Research Notes into Short AI-Generated Video Summaries Designed for TikTok-Style Viewing (View Article)
- Cloudflare Expands Crackdown on AI Crawlers by Blocking Multi-Purpose Bots that Collect Website Data for AI Training and Search (View Article)
- OpenClaw, the Open-Source AI Assistant Compatible with OpenAI Models, Launches Native Apps for Android and iPhone (View Article)
- Microsoft Launches New AI Deployment Company with a $2.5 Billion Investment to Help Organizations Build and Scale Enterprise AI Systems (View Article)
- Meta Quietly Launches Pocket, an AI-Powered Game Creation App that Lets Users Build and Play Games Using Natural Language Prompts (View Article)
- Anthropic Expands Claude into Scientific Research with New Lab Features for Designing, Analyzing, and Managing Biology Experiments (View Article)
- Spark, the AI-Powered Email App, Launches Native Mac AI Features to Help Users Write, Summarize, and Manage Emails More Efficiently (View Article)
- Google Updates AI Search to Highlight Original Recipe Websites More Prominently, Helping Food Creators Regain Traffic Lost to AI Summaries (View Article)
- Microsoft’s Leaked Copilot OS Reveals AI-Powered Operating System Designed to Automate Tasks Across Windows and Applications (View Article)
- Foldimate, the AI-Powered Laundry Folding Robot, Opens Preorders for Its $8,000 Home Robot That Automatically Sorts and Folds Clothes (View Article)
- Anthropic Reportedly Discusses Custom AI Chip Partnership with Samsung to Reduce Reliance on Nvidia Hardware for Claude Models (View Article)
- SpaceX is Reportedly Developing an AI-Powered Personal Device that Could Compete with Smartphones through Voice-First AI Interactions (View Article)
- Acti Brings AI Agents Directly to Smartphone Keyboards, Letting Users Complete Tasks Across Any App Without Switching Applications (View Article)
- Trunk Tools Cuts Construction Document Review From 60 Days to 10 by Using Specialized AI Models Instead of General-Purpose Chatbots (View Article)
- Alibaba Introduces AI Agent Framework that Reduces Token Usage by 99% by Loading Only the Tools Needed for Each Task (View Article)
- Lawsuit Claims ChatGPT 4o Encouraged a Man’s Manic Delusions and Self-Harm, Raising New Questions About AI Safety and Liability (View Article)
- Netflix Revives Wonka’s Golden Ticket Promotion With AI-Powered Interactive Experience Inspired by Gene Wilder’s Classic Film (View Article)
- Square, the Payment and Point-of-Sale Platform Used by Restaurants and Small Businesses, Lets Restaurants Accept Orders Directly from ChatGPT and Claude Through New AI Ordering Integration (View Article)
Work With Me:
If you enjoy this newsletter, please forward it to your friends and colleagues.
Follow me on LinkedIn and X/Twitter to see my latest posts.
Have a wonderful week!
Andrew Bolis
|