model releaseAnthropic

Anthropic releases Fable 5, bringing capabilities of restricted Mythos model to public with $10/$50 per 1M token pricing

TL;DR

Anthropic has released Fable 5, making capabilities from its previously restricted Mythos model available to the public. The company claims Fable 5 beats GPT-5.5, Gemini 3.1 Pro, and its own Opus 4.8 in internal testing, with pricing set at $10 per million input tokens and $50 per million output tokens after a free trial period ending June 22.

2 min read
1

Anthropic releases Fable 5, bringing capabilities of restricted Mythos model to public with $10/$50 per 1M token pricing

Anthropic has released Fable 5, a new model family that brings many capabilities of its restricted Mythos system to the public. The company will charge $10 per million input tokens and $50 per million output tokens after a free trial period ending June 22.

Performance claims and benchmark positioning

According to Anthropic's internal testing, Fable 5 outperforms the company's previous flagship Opus 4.8, as well as OpenAI's GPT-5.5 and Google's Gemini 3.1 Pro. Anthropic states the model shows its largest advantages on longer, more complex tasks.

The company claims Fable 5 excels at software engineering, drug design, and document analysis, while also improving in areas where previous Claude models underperformed, particularly vision tasks.

Vision capabilities demonstration

Anthropic demonstrated Fable 5's vision improvements through gaming benchmarks. The company's previous model, Claude 3.7 Sonnet, required an overlay interface to play Pokémon Red and had to reevaluate positions between frames, essentially operating blind between actions.

Fable 5 can complete Pokémon FireRed (2004 Game Boy Advance) using only vision input, according to Anthropic. The company says this translates to practical capabilities like extracting precise numbers from scientific figures and rebuilding web application source code from screenshots.

Safety measures and routing

Fable 5 includes automatic routing to the less-capable Opus 4.8 for certain topics as a safety measure. Anthropic acknowledges these safeguards are "tuned conservatively" and will sometimes flag harmless requests, triggering in less than 5 percent of sessions on average.

Mythos 5 and Project Glasswing

Anthropic also released Mythos 5, which runs on the same underlying model as Fable 5 but with fewer safeguards. Mythos 5 will initially be available only through Project Glasswing before expanding to a trusted access program.

Mythos, first revealed in early April through Project Glasswing, was shared with select partners including Apple and NVIDIA to help organizations defend against AI-powered cyberattacks. The project prompted discussions about AI regulation at the White House.

Pricing and availability

Claude subscribers can use Fable 5 without spending usage credits until June 22. After that date, pricing is set at $10 per million input tokens and $50 per million output tokens. Anthropic states it aims to restore Fable 5 as a standard part of subscription plans "when sufficient capacity allows."

What this means

The $50 per million output tokens positions Fable 5 as one of the most expensive publicly available models, signaling Anthropic's confidence in its capabilities for high-value tasks. The conservative safety routing and restricted Mythos 5 access suggest the underlying model has capabilities Anthropic considers risky to deploy widely, continuing a pattern of staged releases for frontier models. The vision improvements, if verified independently, represent meaningful progress in multimodal capabilities beyond text-only tasks.

Related Articles

changelog

Anthropic reverses course, makes Claude Fable 5 permanent on subscription plans

Anthropic announced July 18 that Claude Fable 5 will remain available on subscription plans, reversing its previous decision to make the model API-only. Max and Team Premium subscribers will receive access at 50% of standard limits starting July 20, while Pro and Team Standard users get a one-time $100 credit.

model release

Moonshot AI releases 2.8T parameter Kimi K3, pricing at $3/$15 per million tokens

Chinese AI lab Moonshot AI released Kimi K3, a 2.8 trillion parameter model priced at $3 per million input tokens and $15 per million output tokens. The model is currently available via API, with open weights promised by July 27, 2026. This represents the most expensive pricing from a Chinese AI lab to date, matching Anthropic's Claude Sonnet series.

changelog

Anthropic launches rupee pricing for Claude in India at ₹2,000/month, its second-largest market

Anthropic has begun displaying rupee-denominated pricing for Claude subscriptions in India, its second-largest market after the US with 5.8% of global usage. Claude Pro is priced at ₹2,000 ($21) monthly when billed annually, compared to $17 in the US, with Indian prices including local taxes.

changelog

Anthropic extends Claude Fable 5 access through July 19 amid GPT-5.6 Sol competition

Anthropic has extended Claude Fable 5 access on all paid plans through July 19, 2026, marking another extension of the advanced model's availability. The extension comes after OpenAI released GPT-5.6 Sol, which is classified in the same Fable/Mythos model tier.

Comments

Loading...