Follow ZDNET: Add us as a preferred source on Google.
ZDNET’s key takeaways
- Anthropic is releasing Claude Fable 5 for general users.
- Fable 5 uses Mythos-class power with safety controls.
- Pricing is about twice that of Claude Opus 4.8.
Anthropic has announced a defanged version of its fabled (and highly restricted) Mythos large language model. Called Claude Fable 5, the company describes the new AI as “a Mythos-class model made safe for general use.”
Mythos was introduced back in April to great fanfare as a model capable of finding vulnerabilities in code that neither experienced developers nor other AIs could find.
Also: US workers are the world’s biggest AI skeptics – and it’s not just about job loss
Offered as the central component of a Manhattan Project-like team effort called Project Glasswing, Mythos was considered far too dangerous to be allowed to fall into the wrong hands. As such, the model has been made available only to Glasswing partners, including Amazon Web Services, Anthropic, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, Nvidia, and Palo Alto Networks.
Up until now, Mythos was considered a preview product. Now, Anthropic is launching Claude Mythos 5, which will be available to all users with access to Mythos Preview. The company said, “We plan to expand access over time through a more systematic trusted-access program.”
No significant details were provided about any difference between Mythos Preview and Mythos 5, except that the latter appears to be the post-beta version of Mythos.
Also: Apple’s new Siri AI comes with hidden costs that power users should know of
Anthropic also released Fable 5. The company described this technology as the same underlying model as Mythos, but with safeguards. Those safeguards block responses in specific high-risk areas of cybersecurity and biology.
Yeah, the “biology” reference raised red flags. Have they seen biological weapon prompts or responses in their logs? I’ve asked Anthropic. I highly doubt they’ll answer.
Interestingly, if a Fable 5 prompt veers into one of those high-risk areas, the model drops down to Opus 4.8, which has its own restrictions. Ever since version Opus 4.7, Anthropic blocks “Activities that are almost always used maliciously and have little to no legitimate defensive application, such as mass data exfiltration or ransomware code development.”
Professionals with a security clearance from Anthropic can use Opus 4.7 and 4.8 to perform blocked security activities in the course of doing their jobs. Disclosure: I am an authorized member of Anthropic’s Cyber Verification Program, so I have access to these capabilities as part of my cyberwarfare, cyberdefense, and counterterrorism work. It’s not yet clear whether those of us certified through the Cyber Verification Program can perform blocked queries with Fable 5.
Also: I used ChatGPT to build a free PDF editor because I didn’t trust it to change my files – it’s glorious
Anthropic seems confident that Fable can’t be utilized for nefarious purposes. They said, “Early data shows at least 95% of Fable sessions run entirely on Fable’s own responses, with no fallback. We extensively red-teamed our classifiers to test their robustness against jailbreaks. Internally, we ran an external bug bounty that produced no universal jailbreaks in over 1,000 hours of testing. We then worked with external red-teaming orgs which also failed to find universal jailbreaks.”
From the announcement, we don’t have much information to share with you about Fable 5. However, Anthropic did share some customer comments about the release.
An unnamed representative of vibe-coding platform maker Base44 said, “Fable is much deeper and better at one-shotting full apps, and its tool calling is excellent.”
Also: AI can identify intimate partner violence years before people disclose it, but is that safe?
Genspark provides an all-in-one AI workspace. A representative from that company said, “Fable came out #1 on our evals, winning head-to-head against every model we tested. It was significantly stronger on the hardest tasks in the set — UI design and game coding.”
A representative from e-commerce marketplace Rakuten said, “At the highest effort, Fable reflects on and validates its own work. For us, that’s what makes highly autonomous operations possible — the extra thinking pays for itself.”
Speaking of payment, pricing for Fable 5 and the new Mythos 5 release is $10 per million input tokens and $50 per million output tokens. That’s about twice the price of Claude Opus 4.8.
Anthropic has a fairly unique rollout plan for Fable 5, based on their expectation of high demand. Here’s what the company said:
• From today through June 22, Fable 5 is included on Pro, Max, Team, and seat-based Enterprise plans at no extra cost.
• On June 23, we’ll remove Fable 5 from those plans. Using it after that will require usage credits.
• After this point-we aim to restore Fable 5 as a standard part of subscription plans. We intend to do this as quickly as we can.
Anthropic hasn’t provided any deep explanation of its naming choices for Mythos and Fable. According to Britannica, a mythos is a “complex body of sacred narratives, traditional stories, or belief systems that explain the origins of the world and cultural values.” By contrast, a fable is a “short, fictional story designed to teach a specific moral lesson.” Read into that what you wish.
I’m sure we’ll learn more about both of these releases. If my Max plan can host Fable 5, I’ll try and test it out with some coding challenges when it becomes available. So stay tuned.
Even at twice the price of Opus, would you use Claude Fable 5 for coding work if it gave you Mythos-level power with built-in safety fallbacks? Let us know in the comments below.
You can follow my day-to-day project updates on social media. Be sure to subscribe to my weekly update newsletter, and follow me on Twitter/X at @DavidGewirtz, on Facebook at Facebook.com/DavidGewirtz, on Instagram at Instagram.com/DavidGewirtz, on Bluesky at @DavidGewirtz.com, and on YouTube at YouTube.com/DavidGewirtzTV.