- AI Fire
- Posts
- 🎉 Fable & Mythos 5 Return Globally!!!
In partnership with
Plus: Claude Sonnet 5 vs Opus 4.8: Same Result, 50% Cheaper? Read This Before Switching (Free Comparison You Need)
Good news for Claude users: Fable 5 is finally coming back. After weeks of restrictions, Anthropic says global access returns across Claude sites today. Try it while you can.
What's on FIRE 🔥
IN PARTNERSHIP WITH HUBSPOT
Want to get the most out of ChatGPT?
ChatGPT is a superpower if you know how to use it correctly.
Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.
Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.
AI INSIGHTS
As of today, July 1, access to both Fable 5 & Mythos 5 has been officially restored. US gov. has lifted the export controls, bringing an end to the 19-day suspension.
Paid users can access Fable 5 globally on web and mobile right now. Through July 7, you can use up to 50% of your weekly usage limit on it. After that, or if you hit the cap, you'll need usage credits. Keep an eye on your quota, as Fable 5 draws down usage a lot faster than Opus 4.8.
If your prompt hits a safeguard, it gets automatically rerouted to Opus 4.8. Note that a small fraction of routine coding and debugging tasks might trigger a false positive and get bumped to Opus.
On June 12, the models were pulled after Amazon researchers found a "jailbreak" that let Fable 5 bypass safeguards to write exploit code. Because Anthropic couldn't verify user nationality in real-time, they had to shut access down for everyone globally.
Working directly with the US government, Anthropic trained a new classifier that blocks that specific jailbreak technique in over 99% of cases. Government researchers tested the new guardrails and confirmed they are extraordinarily strong.
Mythos 5 model was actually cleared a few days earlier on June 26, but access is currently limited to a specific set of US organizations. Fable 5 is back in action and ready for your heavy lifting, just with tighter security.
PRESENTED BY VIKTOR
Your whole marketing stack, answering in one Slack thread.
Meta in one tab, TikTok in another, Klaviyo and GA4 in two more. Viktor is an AI employee that pulls all of them into a single Slack thread. Ask for blended CAC, yesterday's flow revenue, or the campaign to cut, and get one answer instead of four logins.
AI SOURCES FROM AI FIRE
1. FREE: 5 Claude Connectors I Use to Make My Business Apps Scary Powerful. I tested 100+ Claude connectors & found 5 that instantly transformed my workflows. From automating tasks to smarter data insights, you can use these setups.
2. 5 Proven Claude AI Side Hustles Paying More Than Your Full-Time Job Even in 2027!? 5 actionable ways to for side income that can surpass your day job. Each one is practical, fast to implement, and profitable, not vague ideas.
3. Boost Claude’s Responses: 5 Easy Tricks You Can Use (Apply to New Claude Sonnet 5). Make Claude smarter, faster, and more useful, the kind of answers you didn’t know it could give. There's no trial and error needed.
4. [Claude AI Mastery Playbook] Lesson 9: I Fixed Claude Code’s Memory Issue with 5 Simple Markdown Files. These 5 files can give Claude Code context and project memory. Apply this system to a healthcare app demo in VS Code, and see why reviewing files and plans first keeps your workflow under control.
SONNET 5 VS. OPUS 4.8: WHO WINS
Anthropic posted a "5" made from flowers and leaves on X, and within a few hours it already had 308K views.
→ The announcement was just 2 lines: Sonnet 5 is now the default model for Free and Pro users, available across all Claude apps today.
After reading that, I sat down and did the math, and the number that came out was hard to ignore: same task, Sonnet 5 costs $0.13 while Opus 4.8 costs $0.24.
If you're paying for Opus 4.8 every day, you're paying almost double for results that Sonnet 5 can handle too.
→ So is Sonnet 5 actually as good as Opus 4.8, or is it just cheaper? I tested to see both performance, cost, and practical results so you know whether the cheaper Opus 4.8 actually beats Sonnet 5. Here's what I found out.
TODAY IN AI
AI HIGHLIGHTS
👀 A viral list claims 9 major US tech companies are considering Chinese AI alternatives. The big names make it worth watching. Here's the list if you're curious.
📱 NotebookLM dropped Video Overviews, turning complex topics into engaging 60-second clips with images. Doomscrolling finally becomes useful. See how it works.
🕵️ Claude Code is accused of hiding China proxy and timezone signals inside system prompts. The good news: Anthropic staff says it’s being rolled back. Report link here.
🚀 SpaceX showed investors a handset-like AI phone-ish device, slimmer than an iPhone. The prototype is early, the design may change, and Elon Musk says it’s false.
👉 ICYMI: Here’s a recap of Google’s biggest AI updates from June, including Gemini 3.5 Live Translate, Android 17 and the new Google Home Speaker, built for Gemini.
💸 Palantir CEO Alex Karp criticized OpenAI and Anthropic’s token model. His point: companies are tired of rising AI costs and want control back, something is broken.
☁️ Meta shares jumped nearly 9% after reports it’s building a cloud business to sell excess AI compute, turning huge infrastructure spending into possible new revenue.
💰 Big AI Fundraising: Together AI raised $800M at an $8.3B valuation, up from $3.3B 16 months ago, as demand grows for cheaper AI infrastructure and open-source model hosting.
NEW EMPOWERED AI TOOLS
🧠 Claude Sonnet 5 is Anthropic’s most agentic Sonnet yet, with top-tier intelligence for coding and everyday professional work.
🎥 Gemini Omni Flash supports high-quality video generation & conversational editing from text, image and video inputs.
⌨️ Acti is an agentic keyboard for mobile commands and search. It gets your intent and brings back the right result right where you are.
🤝 Humalike builds the behavioral infrastructure for humanlike AI agents. The social skills & proactiveness your agents have been missing.
🌐 Tabstack Browser Automation automates the web in your app or agent, no browser to host. The result streams back in one API call.
AI BREAKTHROUGH
OpenAI launched GeneBench-Pro, a new benchmark designed to test whether AI models can handle the messy, ambiguous reality of computational biology and genomics.
Unlike standard tests, this benchmark is built to evaluate "research taste", an AI's ability to navigate noisy data, choose the correct analytical path, and make the complex judgment calls required in real-world science.
The suite features 129 complex problems spanning genomics, translational medicine, and quantitative biology. Models must independently figure out how to reach the final answer.
GeneBench-Pro forces models to distinguish between actual biological phenomena and statistical noise as they explore the data.
OpenAI had 82 of the questions reviewed by external domain experts to verify their realism. Human reviewers estimated that a single problem would take a trained human expert 20 to 40 hours to solve.
OpenAI’s GPT-5.6 Sol (in Pro mode) currently leads with a 31.5% pass rate. Claude Opus 4.8 hitting 16.0% and Gemini 3.5 Flash at 8.1%.
The data confirms that while current models are improving rapidly, they are not yet capable of replacing human scientists for full autonomous research. For transparency, OpenAI is open-sourcing a subset of the questions for independent verification.
We read your emails, comments, and poll replies daily
How would you rate today’s newsletter?Your feedback helps us create the best newsletter possible |
Hit reply and say Hello – we'd love to hear from you!
Like what you're reading? Forward it to friends, and they can sign up here.
Cheers,
The AI Fire Team
Reply