Claude Fable 5 vs. GPT-5.5 Pro: 12 tasks compared via the BotHub aggregator

The authors of BotHub's Habr blog pitted two top-tier language models against each other, Anthropic's Claude Fable 5 and OpenAI's GPT-5.5 Pro, on 12 practical tasks. They tested both models through their own neural network API aggregator.

Why this particular pair

In a previous article, the authors compared Claude Opus 4.8, GPT-5.5, and Gemini 3.1 Pro, but noted upfront that GPT-5.5 Pro was left out. The fairer matchup for it is Claude Fable 5, which was unavailable at the time of publication. That changed when Anthropic unexpectedly made Claude Fable 5 publicly available again.

• Claude Fable 5 is the first model in Anthropic's new Mythos class, which sits above Opus in the company's lineup
• Anthropic unexpectedly made Fable 5 publicly available again
• The comparison relies on 12 practical, everyday tasks rather than formal benchmarks
• Testing ran through the BotHub API aggregator, not through the models' web interfaces

How the testing works

The authors deliberately skipped off-the-shelf benchmarks and vendors' marketing claims, betting instead on real everyday tasks. They ran the tests through BotHub, an API-based neural network aggregator. According to the authors, this strips out the crutches and workarounds that quietly help models inside their own branded web interfaces. The setup also shows right away how much each task costs to solve.

Cost is measured in CAPS, BotHub's internal currency, which is tied to the number of tokens consumed. One ruble typically buys between 4000 and 6500 CAPS, depending on the purchase size. For this comparison, the authors used a rate of roughly 1 ruble = 6370 CAPS because their pricing plan had changed. Both models belong to the premium tier, so the per-task price tags came out noticeably higher than in the earlier comparison of Opus 4.8, GPT-5.5, and Gemini 3.1 Pro.

As before, the authors pick the winner of each task largely on subjective grounds, and they say so openly. All testing materials are publicly available, so readers can form their own opinions and choose their own favorite model.

What makes the format interesting

Publishing all the testing materials is part of the authors' method. Instead of simply declaring a winner, they let readers judge both models' answers on each of the 12 tasks and reach their own conclusions, even if those differ from the authors'. This approach matters most for premium models like Claude Fable 5 and GPT-5.5 Pro, where the cost of a request becomes part of the decision. Teams weigh not just the quality of an answer but also its price in CAPS.

What this means

With Claude Fable 5 publicly available again, it is now possible for the first time to compare Anthropic's top Mythos-class model directly with OpenAI's top Pro version of GPT-5.5 on equal API terms, rather than relying on developers' marketing promises. For teams choosing a model for specific workloads, practical comparisons like this one, with transparent token-based cost tracking, are worth more than formal benchmarks, which premium models usually pass almost equally well.

Frequently asked questions

What is Anthropic's Mythos model class?

Mythos is a new class of Anthropic models that sits above Opus in the company's lineup. Claude Fable 5 is the first model released in this class.

How did the comparison calculate the cost of model requests?

Cost was measured in CAPS, the internal currency of the BotHub aggregator, which is tied to the number of tokens consumed. One ruble typically equals 4000-6500 CAPS depending on the purchase size; this test used a rate of roughly 1 ruble = 6370 CAPS.