Anthropic Launches Claude Fable 5.1 and Mythos 5.1, Promising Cheaper, Smarter AI for Coding

Anthropic, an AI company based in the US, has been working to improve its Claude family of large language models. The company previously launched Fable 5, which targeted high-end coding and scientific tasks, but customers raised concerns about cost and strict content safeguards.

As the AI field heated up in 2026, with rivals from China and elsewhere gaining ground, Anthropic faced pressure to deliver faster, cheaper, and more flexible models for enterprise and research use. Customers wanted models that could handle complex, long-running programming work without breaking the bank or running into unnecessary restrictions.

This set the stage for Anthropic's next move: a new generation of models that would address both price and performance complaints, while maintaining security and reliability.

On September 1, 2026, Anthropic officially announced the launch of Claude Fable 5.1 and Claude Mythos 5.1. According to [Bloomberg Technology](https://www.bloomberg.com/news/articles/2026-09-01/anthropic-says-new-fable-5-1-ai-model-is-cheaper-better-at-coding), these models were designed to be both better at coding and significantly more economical than their predecessors. Fable 5.1 is positioned as the main model for the public, while Mythos 5.1 is available only to trusted partners and has fewer restrictions.

TechCrunch reported that Fable 5.1 introduces changes to reduce costs and ease some overly strict safeguards, responding to earlier user feedback. The Verge highlighted that Fable 5.1 can be up to 45 percent cheaper for highly agentic work, with typical workloads seeing a 25 percent cost reduction compared to Fable 5.

The Next Web and Slashdot noted that cache read prices have dropped by 75 percent, making large-scale AI use more affordable for businesses and developers. The models also achieved a notable jump in scientific research benchmarks, with Fable 5.1 scoring 52.6 percent on Terminal-Bench-Science 0.1, compared to 24.7 percent for Fable 5.

Cybersecurity was another focus. GBHackers On Security and Digital Trends explained that both models come with advanced cybersecurity capabilities. Fable 5.1 has new features for secure coding and operations, though it maintains higher safeguards than Mythos 5.1, which is restricted to vetted users due to its looser controls.

Anthropic also introduced new infrastructure for watermarking data outputs, as reported by The New Stack, though developers quickly found that the watermark's visibility can vary depending on the type of content generated.

As of early September 2026, Claude Fable 5.1 is generally available to the public, offering improved coding, lower costs, and enhanced research performance. Mythos 5.1, nearly identical in core technology but less restricted, is limited to trusted partners through Anthropic's access programs.

Early reports from developers and tech news outlets indicate that the upgrades have already begun to impact the market, with some benchmarking results showing clear gains over previous models and international competitors. However, some open questions remain about the consistency of Anthropic's watermarking system and the long-term effectiveness of the security safeguards.

The next steps will likely involve broader adoption of Fable 5.1 among enterprise customers and developers, along with close scrutiny of its cybersecurity features. Industry watchers are paying attention to how Anthropic's pricing changes will influence competition, especially as Chinese open-weight models continue to gain commercial traction. The effectiveness and reliability of the new watermarking technology may also face more testing as users attempt to audit and verify AI-generated outputs.