Thu, 3 Sep

Claude Fable 5.1 Flagship AI Is Temporarily Free to Test

Max Ivanov · 03.09.2026 18:57 · 2 min read

Benchmark platform Arena has opened temporary free access to Anthropic’s top AI model, Claude Fable 5.1. You can try the algorithm in Direct Mode for free for two days — the promotion ends on September 5 at 6:00 PM Kyiv time.

It’s important to note that Anthropic itself hasn’t made the model free on its own resources. In the official web interface and app, access is locked behind paid tiers: Pro, Max, Team, and Enterprise. Developers pay for each API request.

Capabilities and Technical Specs

Anthropic introduced Fable 5.1 in early September, positioning it as its most powerful solution for writing complex code, conducting in-depth research, and handling multi-step agentic tasks.

According to Anthropic’s technical documentation, the system features:

  • A context window of up to 1 million tokens;
  • A maximum single response length of up to 128,000 tokens;
  • An integrated adaptive reasoning mode;
  • Multimodal processing of text and image inputs;
  • Knowledge cutoff of June 2026.

The standard API rate for Fable 5.1 is $10 per million input tokens and $50 per million output tokens. The engineers also cut the cost of reading cached context to $0.25 per million tokens, which can reduce expenses on regular agentic pipelines by nearly half.

Leading the Benchmarks

In the public Arena Leaderboard, the model immediately took the top spot in the overall text AI ranking, while the Fable 5.1 Max variant achieved the highest score in web development tests.

Related: Anthropic Releases Claude Fable 5.1: Doubles Scientific Performance, Cuts Cached Context Costs by 4x, and Expands Context to 1M Tokens.

You can test Claude Fable 5.1 without registration or a subscription on Arena’s direct chat page until the promo period ends.

Enjoy VseZavislo?

Add us to your preferred Google sources to see our news more often.

Add us to your Google

Share

Leave a comment