In recent months, the AI community has been buzzing about a rather mysterious name: Claude Mythos. According to previously revealed information, this is a powerful AI model capable of detecting a large number of security vulnerabilities in software systems. Because of this exceptional capability, Anthropic has implemented a very cautious release process instead of making a public announcement from the outset.
Anthropic has now officially introduced the Claude Fable 5 – the first model in the Mythos family that is accessible to the general public.
According to the company, Fable 5 strikes a balance between performance and safety. It inherits many technologies from the Mythos line but adds control mechanisms to limit the risk of abuse. This allows users to experience one of Anthropic's most advanced AI models without having to participate in special testing programs.
What's new in Claude Fable 5?
Anthropic describes the Fable 5 as the most powerful model the company has ever released for the general public.
This model is optimized for engineering and knowledge processing tasks, especially in the field of software development. Its ability to analyze, understand context, and handle complex tasks is significantly improved compared to previous generations of Claude.
Beyond text, Fable 5 also boasts upgraded image and multimedia data processing capabilities. This allows the AI to better understand charts, technical diagrams, screenshots, or other types of visual documents.
An interesting detail shared by Anthropic is that Fable 5 completed the Pokémon FireRed game on the Game Boy Advance using computer vision, something many previous models had failed to do. While this isn't an official measure of AI capability, it does demonstrate a significant improvement in the model's ability to observe, analyze images, and make decisions.
How do Fable 5 and Mythos 5 differ?
This is a question that many people are interested in. Basically, Claude Fable 5 is not the most powerful version in the Mythos family.
Anthropic stated that the new Claude Mythos 5 is the model that possesses all the cybersecurity research and analysis capabilities the company previously mentioned. However, due to potential security risks, Mythos 5 is currently only available to a small group of trusted cybersecurity experts and software engineers through the Project Glasswing program.
Fable 5 is considered a "safer" version of Mythos 5. In other words, users still have access to many of the advanced capabilities of the Mythos family, but within a more tightly controlled scope.
Is the Claude Fable 5 safe?
Security is one of the themes Anthropic emphasized most when introducing Fable 5. According to internal testing results, this model was designed to reject requests related to cyberattack planning, vulnerability development, or attempts to bypass security defenses.
Anthropic stated that Fable 5 still maintains its ability to reject jailbreaks even when users employ various common jailbreak techniques to trick the system.
Beyond cybersecurity, Anthropic is also being particularly cautious about topics related to biology and chemistry.
Currently, most in-depth questions in these two areas will not be handled directly by Fable 5. Instead, the system will transition to a different model, Claude Opus 4.8, to ensure better risk control.
In addition, Anthropic stated that they are still working with regulatory agencies in the US during the evaluation and preparation for future releases of the Mythos series.
How to use Claude Fable 5
Claude's paying users can now start experiencing Fable 5. However, Anthropic says the rollout will be in phases rather than being fully available all at once.
Initially, Fable 5 was only available through paid subscription plans, including:
- Claude Pro
- Claude Max
- Claude Team
- Claude Enterprise
One important point to note is that Fable 5 will only appear in the model selection list until June 22nd. After that date, the model will be removed from the default list. This does not mean users will lose access entirely.
Note that after the initial rollout phase, Fable 5 will switch to a Usage Credits billing system. This is a credit system based on actual usage, allowing users to continue accessing the service even after exceeding their subscription's usual usage limit.
Some business customers may have been pre-allocated a certain amount of Usage Credits. However, if they exceed their allocated limit, additional charges will appear on their next invoice.
In other words, Fable 5 will become a premium resource rather than a free option included in a standard subscription package.
Claude Fable 5 developer pricing
Anthropic has also opened up access to Fable 5 through its developer API.
The prices are as follows:
This price is relatively high compared to many current AI models, reflecting the large amount of computing resources that Fable 5 requires.
The biggest reason lies in processing power and operating costs. According to Anthropic, Fable 5 consumes twice as much computing resources as the current Claude Opus series.
While demand for the new model is expected to be high, the company's hardware infrastructure still has certain limitations. Deploying in phases helps Anthropic maintain the stability of the entire Claude platform instead of overloading the system.
The company says its long-term goal is to make Fable 5 a full-fledged part of paid subscription packages. However, no specific timeline for this plan has been given yet.
Is Claude Fable 5 worth experiencing?
For those who frequently use AI for programming, research, or professional tasks, the Fable 5 is one of the most noteworthy models available today.
Being built on Mythos technology gives the model very high reasoning and complex task processing capabilities. At the same time, added safety mechanisms help mitigate the risks that Anthropic was concerned about when it was widely released.
Although not yet a full version of Mythos, Claude Fable 5 can be seen as the first opportunity for mainstream users to access the new generation of AI that Anthropic has been quietly developing.
Claude Fable 5 marks a new step in Anthropic's AI development strategy. Instead of focusing solely on conversational chatbots, the company is moving towards models capable of handling more specialized tasks in the fields of engineering, research, and knowledge.
The limited release, the application of a separate pricing mechanism, and the phased rollout show that Anthropic is taking a very cautious approach to the Mythos genre.
However, for users who want to explore the latest AI capabilities, the Claude Fable 5 is certainly one of the most worthwhile models to experience this year.